prometheus/docs/querying/functions.md

837 lines
37 KiB
Markdown
Raw Normal View History

---
title: Query functions
nav_title: Functions
sort_rank: 3
---
# Functions
Some functions have default arguments, e.g. `year(v=vector(time())
instant-vector)`. This means that there is one argument `v` which is an instant
vector, which if not provided it will default to the value of the expression
`vector(time())`.
_Notes about the experimental native histograms:_
* Ingesting native histograms has to be enabled via a [feature
flag](../feature_flags.md#native-histograms). As long as no native histograms
have been ingested into the TSDB, all functions will behave as usual.
* Functions that do not explicitly mention native histograms in their
promql: Separate `Point` into `FPoint` and `HPoint` In other words: Instead of having a “polymorphous” `Point` that can either contain a float value or a histogram value, use an `FPoint` for floats and an `HPoint` for histograms. This seemingly small change has a _lot_ of repercussions throughout the codebase. The idea here is to avoid the increase in size of `Point` arrays that happened after native histograms had been added. The higher-level data structures (`Sample`, `Series`, etc.) are still “polymorphous”. The same idea could be applied to them, but at each step the trade-offs needed to be evaluated. The idea with this change is to do the minimum necessary to get back to pre-histogram performance for functions that do not touch histograms. Here are comparisons for the `changes` function. The test data doesn't include histograms yet. Ideally, there would be no change in the benchmark result at all. First runtime v2.39 compared to directly prior to this commit: ``` name old time/op new time/op delta RangeQuery/expr=changes(a_one[1d]),steps=1-16 391µs ± 2% 542µs ± 1% +38.58% (p=0.000 n=9+8) RangeQuery/expr=changes(a_one[1d]),steps=10-16 452µs ± 2% 617µs ± 2% +36.48% (p=0.000 n=10+10) RangeQuery/expr=changes(a_one[1d]),steps=100-16 1.12ms ± 1% 1.36ms ± 2% +21.58% (p=0.000 n=8+10) RangeQuery/expr=changes(a_one[1d]),steps=1000-16 7.83ms ± 1% 8.94ms ± 1% +14.21% (p=0.000 n=10+10) RangeQuery/expr=changes(a_ten[1d]),steps=1-16 2.98ms ± 0% 3.30ms ± 1% +10.67% (p=0.000 n=9+10) RangeQuery/expr=changes(a_ten[1d]),steps=10-16 3.66ms ± 1% 4.10ms ± 1% +11.82% (p=0.000 n=10+10) RangeQuery/expr=changes(a_ten[1d]),steps=100-16 10.5ms ± 0% 11.8ms ± 1% +12.50% (p=0.000 n=8+10) RangeQuery/expr=changes(a_ten[1d]),steps=1000-16 77.6ms ± 1% 87.4ms ± 1% +12.63% (p=0.000 n=9+9) RangeQuery/expr=changes(a_hundred[1d]),steps=1-16 30.4ms ± 2% 32.8ms ± 1% +8.01% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=10-16 37.1ms ± 2% 40.6ms ± 2% +9.64% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=100-16 105ms ± 1% 117ms ± 1% +11.69% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=1000-16 783ms ± 3% 876ms ± 1% +11.83% (p=0.000 n=9+10) ``` And then runtime v2.39 compared to after this commit: ``` name old time/op new time/op delta RangeQuery/expr=changes(a_one[1d]),steps=1-16 391µs ± 2% 547µs ± 1% +39.84% (p=0.000 n=9+8) RangeQuery/expr=changes(a_one[1d]),steps=10-16 452µs ± 2% 616µs ± 2% +36.15% (p=0.000 n=10+10) RangeQuery/expr=changes(a_one[1d]),steps=100-16 1.12ms ± 1% 1.26ms ± 1% +12.20% (p=0.000 n=8+10) RangeQuery/expr=changes(a_one[1d]),steps=1000-16 7.83ms ± 1% 7.95ms ± 1% +1.59% (p=0.000 n=10+8) RangeQuery/expr=changes(a_ten[1d]),steps=1-16 2.98ms ± 0% 3.38ms ± 2% +13.49% (p=0.000 n=9+10) RangeQuery/expr=changes(a_ten[1d]),steps=10-16 3.66ms ± 1% 4.02ms ± 1% +9.80% (p=0.000 n=10+9) RangeQuery/expr=changes(a_ten[1d]),steps=100-16 10.5ms ± 0% 10.8ms ± 1% +3.08% (p=0.000 n=8+10) RangeQuery/expr=changes(a_ten[1d]),steps=1000-16 77.6ms ± 1% 78.1ms ± 1% +0.58% (p=0.035 n=9+10) RangeQuery/expr=changes(a_hundred[1d]),steps=1-16 30.4ms ± 2% 33.5ms ± 4% +10.18% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=10-16 37.1ms ± 2% 40.0ms ± 1% +7.98% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=100-16 105ms ± 1% 107ms ± 1% +1.92% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=1000-16 783ms ± 3% 775ms ± 1% -1.02% (p=0.019 n=9+9) ``` In summary, the runtime doesn't really improve with this change for queries with just a few steps. For queries with many steps, this commit essentially reinstates the old performance. This is good because the many-step queries are the one that matter most (longest absolute runtime). In terms of allocations, though, this commit doesn't make a dent at all (numbers not shown). The reason is that most of the allocations happen in the sampleRingIterator (in the storage package), which has to be addressed in a separate commit. Signed-off-by: beorn7 <beorn@grafana.com>
2022-10-28 14:58:40 +00:00
documentation (see below) will ignore histogram samples.
* Functions that do already act on native histograms might still change their
behavior in the future.
* If a function requires the same bucket layout between multiple native
histograms it acts on, it will automatically convert them
appropriately. (With the currently supported bucket schemas, that's always
possible.)
## `abs()`
`abs(v instant-vector)` returns the input vector with all sample values converted to
their absolute value.
## `absent()`
`absent(v instant-vector)` returns an empty vector if the vector passed to it
has any elements (floats or native histograms) and a 1-element vector with the
value 1 if the vector passed to it has no elements.
This is useful for alerting on when no time series exist for a given metric name
and label combination.
```
absent(nonexistent{job="myjob"})
# => {job="myjob"}
absent(nonexistent{job="myjob",instance=~".*"})
# => {job="myjob"}
absent(sum(nonexistent{job="myjob"}))
# => {}
```
In the first two examples, `absent()` tries to be smart about deriving labels
of the 1-element output vector from the input vector.
## `absent_over_time()`
`absent_over_time(v range-vector)` returns an empty vector if the range vector
passed to it has any elements (floats or native histograms) and a 1-element
vector with the value 1 if the range vector passed to it has no elements.
This is useful for alerting on when no time series exist for a given metric name
and label combination for a certain amount of time.
```
absent_over_time(nonexistent{job="myjob"}[1h])
# => {job="myjob"}
absent_over_time(nonexistent{job="myjob",instance=~".*"}[1h])
# => {job="myjob"}
absent_over_time(sum(nonexistent{job="myjob"})[1h:])
# => {}
```
In the first two examples, `absent_over_time()` tries to be smart about deriving
labels of the 1-element output vector from the input vector.
## `ceil()`
`ceil(v instant-vector)` rounds the sample values of all elements in `v` up to
the nearest integer value greater than or equal to v.
* `ceil(+Inf) = +Inf`
* `ceil(±0) = ±0`
* `ceil(1.49) = 2.0`
* `ceil(1.78) = 2.0`
## `changes()`
For each input time series, `changes(v range-vector)` returns the number of
times its value has changed within the provided time range as an instant
vector.
## `clamp()`
`clamp(v instant-vector, min scalar, max scalar)`
clamps the sample values of all elements in `v` to have a lower limit of `min` and an upper limit of `max`.
Special cases:
* Return an empty vector if `min > max`
* Return `NaN` if `min` or `max` is `NaN`
## `clamp_max()`
`clamp_max(v instant-vector, max scalar)` clamps the sample values of all
elements in `v` to have an upper limit of `max`.
## `clamp_min()`
`clamp_min(v instant-vector, min scalar)` clamps the sample values of all
elements in `v` to have a lower limit of `min`.
## `day_of_month()`
`day_of_month(v=vector(time()) instant-vector)` returns the day of the month
for each of the given times in UTC. Returned values are from 1 to 31.
## `day_of_week()`
`day_of_week(v=vector(time()) instant-vector)` returns the day of the week for
each of the given times in UTC. Returned values are from 0 to 6, where 0 means
Sunday etc.
## `day_of_year()`
`day_of_year(v=vector(time()) instant-vector)` returns the day of the year for
each of the given times in UTC. Returned values are from 1 to 365 for non-leap years,
and 1 to 366 in leap years.
## `days_in_month()`
`days_in_month(v=vector(time()) instant-vector)` returns number of days in the
month for each of the given times in UTC. Returned values are from 28 to 31.
## `delta()`
`delta(v range-vector)` calculates the difference between the
first and last value of each time series element in a range vector `v`,
returning an instant vector with the given deltas and equivalent labels.
The delta is extrapolated to cover the full time range as specified in
the range vector selector, so that it is possible to get a non-integer
result even if the sample values are all integers.
The following example expression returns the difference in CPU temperature
between now and 2 hours ago:
```
delta(cpu_temp_celsius{host="zeus"}[2h])
```
`delta` acts on native histograms by calculating a new histogram where each
component (sum and count of observations, buckets) is the difference between
the respective component in the first and last native histogram in
`v`. However, each element in `v` that contains a mix of float and native
histogram samples within the range, will be missing from the result vector.
`delta` should only be used with gauges and native histograms where the
components behave like gauges (so-called gauge histograms).
## `deriv()`
`deriv(v range-vector)` calculates the per-second derivative of the time series in a range
vector `v`, using [simple linear regression](https://en.wikipedia.org/wiki/Simple_linear_regression).
The range vector must have at least two samples in order to perform the calculation. When `+Inf` or
`-Inf` are found in the range vector, the slope and offset value calculated will be `NaN`.
`deriv` should only be used with gauges.
## `exp()`
`exp(v instant-vector)` calculates the exponential function for all elements in `v`.
Special cases are:
* `Exp(+Inf) = +Inf`
* `Exp(NaN) = NaN`
## `floor()`
`floor(v instant-vector)` rounds the sample values of all elements in `v` down
to the nearest integer value smaller than or equal to v.
* `floor(+Inf) = +Inf`
* `floor(±0) = ±0`
* `floor(1.49) = 1.0`
* `floor(1.78) = 1.0`
## `histogram_avg()`
_This function only acts on native histograms, which are an experimental
feature. The behavior of this function may change in future versions of
Prometheus, including its removal from PromQL._
`histogram_avg(v instant-vector)` returns the arithmetic average of observed values stored in
a native histogram. Samples that are not native histograms are ignored and do
not show up in the returned vector.
Use `histogram_avg` as demonstrated below to compute the average request duration
over a 5-minute window from a native histogram:
histogram_avg(rate(http_request_duration_seconds[5m]))
Which is equivalent to the following query:
histogram_sum(rate(http_request_duration_seconds[5m]))
/
histogram_count(rate(http_request_duration_seconds[5m]))
## `histogram_count()` and `histogram_sum()`
_Both functions only act on native histograms, which are an experimental
feature. The behavior of these functions may change in future versions of
Prometheus, including their removal from PromQL._
`histogram_count(v instant-vector)` returns the count of observations stored in
a native histogram. Samples that are not native histograms are ignored and do
not show up in the returned vector.
Similarly, `histogram_sum(v instant-vector)` returns the sum of observations
stored in a native histogram.
Use `histogram_count` in the following way to calculate a rate of observations
(in this case corresponding to “requests per second”) from a native histogram:
histogram_count(rate(http_request_duration_seconds[10m]))
## `histogram_fraction()`
_This function only acts on native histograms, which are an experimental
feature. The behavior of this function may change in future versions of
Prometheus, including its removal from PromQL._
For a native histogram, `histogram_fraction(lower scalar, upper scalar, v
instant-vector)` returns the estimated fraction of observations between the
provided lower and upper values. Samples that are not native histograms are
ignored and do not show up in the returned vector.
For example, the following expression calculates the fraction of HTTP requests
over the last hour that took 200ms or less:
histogram_fraction(0, 0.2, rate(http_request_duration_seconds[1h]))
The error of the estimation depends on the resolution of the underlying native
histogram and how closely the provided boundaries are aligned with the bucket
boundaries in the histogram.
`+Inf` and `-Inf` are valid boundary values. For example, if the histogram in
the expression above included negative observations (which shouldn't be the
case for request durations), the appropriate lower boundary to include all
observations less than or equal 0.2 would be `-Inf` rather than `0`.
Whether the provided boundaries are inclusive or exclusive is only relevant if
the provided boundaries are precisely aligned with bucket boundaries in the
underlying native histogram. In this case, the behavior depends on the schema
definition of the histogram. The currently supported schemas all feature
inclusive upper boundaries and exclusive lower boundaries for positive values
(and vice versa for negative values). Without a precise alignment of
boundaries, the function uses linear interpolation to estimate the
fraction. With the resulting uncertainty, it becomes irrelevant if the
boundaries are inclusive or exclusive.
## `histogram_quantile()`
`histogram_quantile(φ scalar, b instant-vector)` calculates the φ-quantile (0 ≤
φ ≤ 1) from a [classic
histogram](https://prometheus.io/docs/concepts/metric_types/#histogram) or from
a native histogram. (See [histograms and
summaries](https://prometheus.io/docs/practices/histograms) for a detailed
explanation of φ-quantiles and the usage of the (classic) histogram metric
type in general.)
_Note that native histograms are an experimental feature. The behavior of this
function when dealing with native histograms may change in future versions of
Prometheus._
The float samples in `b` are considered the counts of observations in each
bucket of one or more classic histograms. Each float sample must have a label
`le` where the label value denotes the inclusive upper bound of the bucket.
(Float samples without such a label are silently ignored.) The other labels and
the metric name are used to identify the buckets belonging to each classic
histogram. The [histogram metric
type](https://prometheus.io/docs/concepts/metric_types/#histogram)
automatically provides time series with the `_bucket` suffix and the
appropriate labels.
The native histogram samples in `b` are treated each individually as a separate
histogram to calculate the quantile from.
As long as no naming collisions arise, `b` may contain a mix of classic
and native histograms.
Use the `rate()` function to specify the time window for the quantile
calculation.
Example: A histogram metric is called `http_request_duration_seconds` (and
therefore the metric name for the buckets of a classic histogram is
`http_request_duration_seconds_bucket`). To calculate the 90th percentile of request
durations over the last 10m, use the following expression in case
`http_request_duration_seconds` is a classic histogram:
histogram_quantile(0.9, rate(http_request_duration_seconds_bucket[10m]))
For a native histogram, use the following expression instead:
histogram_quantile(0.9, rate(http_request_duration_seconds[10m]))
The quantile is calculated for each label combination in
`http_request_duration_seconds`. To aggregate, use the `sum()` aggregator
around the `rate()` function. Since the `le` label is required by
`histogram_quantile()` to deal with classic histograms, it has to be
included in the `by` clause. The following expression aggregates the 90th
percentile by `job` for classic histograms:
histogram_quantile(0.9, sum by (job, le) (rate(http_request_duration_seconds_bucket[10m])))
When aggregating native histograms, the expression simplifies to:
histogram_quantile(0.9, sum by (job) (rate(http_request_duration_seconds[10m])))
To aggregate all classic histograms, specify only the `le` label:
histogram_quantile(0.9, sum by (le) (rate(http_request_duration_seconds_bucket[10m])))
With native histograms, aggregating everything works as usual without any `by` clause:
histogram_quantile(0.9, sum(rate(http_request_duration_seconds[10m])))
promql(native histograms): Introduce exponential interpolation The linear interpolation (assuming that observations are uniformly distributed within a bucket) is a solid and simple assumption in lack of any other information. However, the exponential bucketing used by standard schemas of native histograms has been chosen to cover the whole range of observations in a way that bucket populations are spread out over buckets in a reasonably way for typical distributions encountered in real-world scenarios. This is the origin of the idea implemented here: If we divide a given bucket into two (or more) smaller exponential buckets, we "most naturally" expect that the samples in the original buckets will split among those smaller buckets in a more or less uniform fashion. With this assumption, we end up with an "exponential interpolation", which therefore appears to be a better match for histograms with exponential bucketing. This commit leaves the linear interpolation in place for NHCB, but changes the interpolation for exponential native histograms to exponential. This affects `histogram_quantile` and `histogram_fraction` (because the latter is more or less the inverse of the former). The zero bucket has to be treated specially because the assumption above would lead to an "interpolation to zero" (the bucket density approaches infinity around zero, and with the postulated uniform usage of buckets, we would end up with an estimate of zero for all quantiles ending up in the zero bucket). We simply fall back to linear interpolation within the zero bucket. At the same time, this commit makes the call to stick with the assumption that the zero bucket only contains positive observations for native histograms without negative buckets (and vice versa). (This is an assumption relevant for interpolation. It is a mostly academic point, as the zero bucket is supposed to be very small anyway. However, in cases where it _is_ relevantly broad, the assumption helps a lot in practice.) This commit also updates and completes the documentation to match both details about interpolation. As a more high level note: The approach here attempts to strike a balance between a more simplistic approach without any assumption, and a more involved approach with more sophisticated assumptions. I will shortly describe both for reference: The "zero assumption" approach would be to not interpolate at all, but _always_ return the harmonic mean of the bucket boundaries of the bucket the quantile ends up in. This has the advantage of minimizing the maximum possible relative error of the quantile estimation. (Depending on the exact definition of the relative error of an estimation, there is also an argument to return the arithmetic mean of the bucket boundaries.) While limiting the maximum possible relative error is a good property, this approach would throw away the information if a quantile is closer to the upper or lower end of the population within a bucket. This can be valuable trending information in a dashboard. With any kind of interpolation, the maximum possible error of a quantile estimation increases to the full width of a bucket (i.e. it more than doubles for the harmonic mean approach, and precisely doubles for the arithmetic mean approach). However, in return the _expectation value_ of the error decreases. The increase of the theoretical maximum only has practical relevance for pathologic distributions. For example, if there are thousand observations within a bucket, they could _all_ be at the upper bound of the bucket. If the quantile calculation picks the 1st observation in the bucket as the relevant one, an interpolation will yield a value close to the lower bucket boundary, while the true quantile value is close to the upper boundary. The "fancy interpolation" approach would be one that analyses the _actual_ distribution of samples in the histogram. A lot of statistics could be applied based on the information we have available in the histogram. This would include the population of neighboring (or even all) buckets in the histogram. In general, the resolution of a native histogram should be quite high, and therefore, those "fancy" approaches would increase the computational cost quite a bit with very little practical benefits (i.e. just tiny corrections of the estimated quantile value). The results are also much harder to reason with. Signed-off-by: beorn7 <beorn@grafana.com>
2024-08-15 12:19:16 +00:00
In the (common) case that a quantile value does not coincide with a bucket
boundary, the `histogram_quantile()` function interpolates the quantile value
within the bucket the quantile value falls into. For classic histograms, for
native histograms with custom bucket boundaries, and for the zero bucket of
other native histograms, it assumes a uniform distribution of observations
within the bucket (also called _linear interpolation_). For the
non-zero-buckets of native histograms with a standard exponential bucketing
schema, the interpolation is done under the assumption that the samples within
the bucket are distributed in a way that they would uniformly populate the
buckets in a hypothetical histogram with higher resolution. (This is also
called _exponential interpolation_.)
If `b` has 0 observations, `NaN` is returned. For φ < 0, `-Inf` is
returned. For φ > 1, `+Inf` is returned. For φ = `NaN`, `NaN` is returned.
promql(native histograms): Introduce exponential interpolation The linear interpolation (assuming that observations are uniformly distributed within a bucket) is a solid and simple assumption in lack of any other information. However, the exponential bucketing used by standard schemas of native histograms has been chosen to cover the whole range of observations in a way that bucket populations are spread out over buckets in a reasonably way for typical distributions encountered in real-world scenarios. This is the origin of the idea implemented here: If we divide a given bucket into two (or more) smaller exponential buckets, we "most naturally" expect that the samples in the original buckets will split among those smaller buckets in a more or less uniform fashion. With this assumption, we end up with an "exponential interpolation", which therefore appears to be a better match for histograms with exponential bucketing. This commit leaves the linear interpolation in place for NHCB, but changes the interpolation for exponential native histograms to exponential. This affects `histogram_quantile` and `histogram_fraction` (because the latter is more or less the inverse of the former). The zero bucket has to be treated specially because the assumption above would lead to an "interpolation to zero" (the bucket density approaches infinity around zero, and with the postulated uniform usage of buckets, we would end up with an estimate of zero for all quantiles ending up in the zero bucket). We simply fall back to linear interpolation within the zero bucket. At the same time, this commit makes the call to stick with the assumption that the zero bucket only contains positive observations for native histograms without negative buckets (and vice versa). (This is an assumption relevant for interpolation. It is a mostly academic point, as the zero bucket is supposed to be very small anyway. However, in cases where it _is_ relevantly broad, the assumption helps a lot in practice.) This commit also updates and completes the documentation to match both details about interpolation. As a more high level note: The approach here attempts to strike a balance between a more simplistic approach without any assumption, and a more involved approach with more sophisticated assumptions. I will shortly describe both for reference: The "zero assumption" approach would be to not interpolate at all, but _always_ return the harmonic mean of the bucket boundaries of the bucket the quantile ends up in. This has the advantage of minimizing the maximum possible relative error of the quantile estimation. (Depending on the exact definition of the relative error of an estimation, there is also an argument to return the arithmetic mean of the bucket boundaries.) While limiting the maximum possible relative error is a good property, this approach would throw away the information if a quantile is closer to the upper or lower end of the population within a bucket. This can be valuable trending information in a dashboard. With any kind of interpolation, the maximum possible error of a quantile estimation increases to the full width of a bucket (i.e. it more than doubles for the harmonic mean approach, and precisely doubles for the arithmetic mean approach). However, in return the _expectation value_ of the error decreases. The increase of the theoretical maximum only has practical relevance for pathologic distributions. For example, if there are thousand observations within a bucket, they could _all_ be at the upper bound of the bucket. If the quantile calculation picks the 1st observation in the bucket as the relevant one, an interpolation will yield a value close to the lower bucket boundary, while the true quantile value is close to the upper boundary. The "fancy interpolation" approach would be one that analyses the _actual_ distribution of samples in the histogram. A lot of statistics could be applied based on the information we have available in the histogram. This would include the population of neighboring (or even all) buckets in the histogram. In general, the resolution of a native histogram should be quite high, and therefore, those "fancy" approaches would increase the computational cost quite a bit with very little practical benefits (i.e. just tiny corrections of the estimated quantile value). The results are also much harder to reason with. Signed-off-by: beorn7 <beorn@grafana.com>
2024-08-15 12:19:16 +00:00
Special cases for classic histograms:
* If `b` contains fewer than two buckets, `NaN` is returned.
* The highest bucket must have an upper bound of `+Inf`. (Otherwise, `NaN` is
returned.)
* If a quantile is located in the highest bucket, the upper bound of the second
highest bucket is returned.
* The lower limit of the lowest bucket is assumed to be 0 if the upper bound of
that bucket is greater than 0. In that case, the usual linear interpolation
is applied within that bucket. Otherwise, the upper bound of the lowest
bucket is returned for quantiles located in the lowest bucket.
Special cases for native histograms (relevant for the exact interpolation
happening within the zero bucket):
* A zero bucket with finite width is assumed to contain no negative
observations if the histogram has observations in positive buckets, but none
in negative buckets.
* A zero bucket with finite width is assumed to contain no positive
observations if the histogram has observations in negative buckets, but none
in positive buckets.
You can use `histogram_quantile(0, v instant-vector)` to get the estimated
minimum value stored in a histogram.
You can use `histogram_quantile(1, v instant-vector)` to get the estimated
maximum value stored in a histogram.
Buckets of classic histograms are cumulative. Therefore, the following should
always be the case:
* The counts in the buckets are monotonically increasing (strictly
non-decreasing).
* A lack of observations between the upper limits of two consecutive buckets
results in equal counts in those two buckets.
However, floating point precision issues (e.g. small discrepancies introduced
by computing of buckets with `sum(rate(...))`) or invalid data might violate
these assumptions. In that case, `histogram_quantile` would be unable to return
meaningful results. To mitigate the issue, `histogram_quantile` assumes that
tiny relative differences between consecutive buckets are happening because of
floating point precision errors and ignores them. (The threshold to ignore a
difference between two buckets is a trillionth (1e-12) of the sum of both
buckets.) Furthermore, if there are non-monotonic bucket counts even after this
adjustment, they are increased to the value of the previous buckets to enforce
monotonicity. The latter is evidence for an actual issue with the input data
and is therefore flagged with an informational annotation reading `input to
histogram_quantile needed to be fixed for monotonicity`. If you encounter this
annotation, you should find and remove the source of the invalid data.
## `histogram_stddev()` and `histogram_stdvar()`
_Both functions only act on native histograms, which are an experimental
feature. The behavior of these functions may change in future versions of
Prometheus, including their removal from PromQL._
`histogram_stddev(v instant-vector)` returns the estimated standard deviation
of observations in a native histogram, based on the geometric mean of the buckets
where the observations lie. Samples that are not native histograms are ignored and
do not show up in the returned vector.
Similarly, `histogram_stdvar(v instant-vector)` returns the estimated standard
variance of observations in a native histogram.
## `double_exponential_smoothing()`
**This function has to be enabled via the [feature flag](../feature_flags.md#experimental-promql-functions) `--enable-feature=promql-experimental-functions`.**
`double_exponential_smoothing(v range-vector, sf scalar, tf scalar)` produces a smoothed value
for time series based on the range in `v`. The lower the smoothing factor `sf`,
the more importance is given to old data. The higher the trend factor `tf`, the
more trends in the data is considered. Both `sf` and `tf` must be between 0 and
1.
For additional details, refer to [NIST Engineering Statistics Handbook](https://www.itl.nist.gov/div898/handbook/pmc/section4/pmc433.htm).
In Prometheus V2 this function was called `holt_winters`. This caused confusion
since the Holt-Winters method usually refers to triple exponential smoothing.
Double exponential smoothing as implemented here is also referred to as "Holt
Linear".
`double_exponential_smoothing` should only be used with gauges.
## `hour()`
`hour(v=vector(time()) instant-vector)` returns the hour of the day
for each of the given times in UTC. Returned values are from 0 to 23.
## `idelta()`
`idelta(v range-vector)` calculates the difference between the last two samples
in the range vector `v`, returning an instant vector with the given deltas and
equivalent labels.
`idelta` should only be used with gauges.
## `increase()`
`increase(v range-vector)` calculates the increase in the
time series in the range vector. Breaks in monotonicity (such as counter
resets due to target restarts) are automatically adjusted for. The
increase is extrapolated to cover the full time range as specified
in the range vector selector, so that it is possible to get a
non-integer result even if a counter increases only by integer
increments.
The following example expression returns the number of HTTP requests as measured
over the last 5 minutes, per time series in the range vector:
```
increase(http_requests_total{job="api-server"}[5m])
```
`increase` acts on native histograms by calculating a new histogram where each
component (sum and count of observations, buckets) is the increase between
the respective component in the first and last native histogram in
`v`. However, each element in `v` that contains a mix of float and native
histogram samples within the range, will be missing from the result vector.
`increase` should only be used with counters and native histograms where the
components behave like counters. It is syntactic sugar for `rate(v)` multiplied
by the number of seconds under the specified time range window, and should be
used primarily for human readability. Use `rate` in recording rules so that
increases are tracked consistently on a per-second basis.
## `info()` (experimental)
_The `info` function is an experiment to improve UX
around including labels from [info metrics](https://grafana.com/blog/2021/08/04/how-to-use-promql-joins-for-more-effective-queries-of-prometheus-metrics-at-scale/#info-metrics).
The behavior of this function may change in future versions of Prometheus,
including its removal from PromQL. `info` has to be enabled via the
[feature flag](../feature_flags.md#experimental-promql-functions) `--enable-feature=promql-experimental-functions`._
`info(v instant-vector, [data-label-selector instant-vector])` finds, for each time
series in `v`, all info series with matching _identifying_ labels (more on
this later), and adds the union of their _data_ (i.e., non-identifying) labels
to the time series. The second argument `data-label-selector` is optional.
It is not a real instant vector, but uses a subset of its syntax.
It must start and end with curly braces (`{ ... }`) and may only contain label matchers.
The label matchers are used to constrain which info series to consider
and which data labels to add to `v`.
Identifying labels of an info series are the subset of labels that uniquely
identify the info series. The remaining labels are considered
_data labels_ (also called non-identifying). (Note that Prometheus's concept
of time series identity always includes _all_ the labels. For the sake of the `info`
function, we “logically” define info series identity in a different way than
in the conventional Prometheus view.) The identifying labels of an info series
are used to join it to regular (non-info) series, i.e. those series that have
the same labels as the identifying labels of the info series. The data labels, which are
the ones added to the regular series by the `info` function, effectively encode
metadata key value pairs. (This implies that a change in the data labels
in the conventional Prometheus view constitutes the end of one info series and
the beginning of a new info series, while the “logical” view of the `info` function is
that the same info series continues to exist, just with different “data”.)
The conventional approach of adding data labels is sometimes called a “join query”,
as illustrated by the following example:
```
rate(http_server_request_duration_seconds_count[2m])
* on (job, instance) group_left (k8s_cluster_name)
target_info
```
The core of the query is the expression `rate(http_server_request_duration_seconds_count[2m])`.
But to add data labels from an info metric, the user has to use elaborate
(and not very obvious) syntax to specify which info metric to use (`target_info`), what the
identifying labels are (`on (job, instance)`), and which data labels to add
(`group_left (k8s_cluster_name)`).
This query is not only verbose and hard to write, it might also run into an “identity crisis”:
If any of the data labels of `target_info` changes, Prometheus sees that as a change of series
(as alluded to above, Prometheus just has no native concept of non-identifying labels).
If the old `target_info` series is not properly marked as stale (which can happen with certain ingestion paths),
the query above will fail for up to 5m (the lookback delta) because it will find a conflicting
match with both the old and the new version of `target_info`.
The `info` function not only resolves this conflict in favor of the newer series, it also simplifies the syntax
because it knows about the available info series and what their identifying labels are. The example query
looks like this with the `info` function:
```
info(
rate(http_server_request_duration_seconds_count[2m]),
{k8s_cluster_name=~".+"}
)
```
The common case of adding _all_ data labels can be achieved by
omitting the 2nd argument of the `info` function entirely, simplifying
the example even more:
```
info(rate(http_server_request_duration_seconds_count[2m]))
```
While `info` normally automatically finds all matching info series, it's possible to
restrict them by providing a `__name__` label matcher, e.g.
`{__name__="target_info"}`.
### Limitations
In its current iteration, `info` defaults to considering only info series with
the name `target_info`. It also assumes that the identifying info series labels are
`instance` and `job`. `info` does support other info series names however, through
`__name__` label matchers. E.g., one can explicitly say to consider both
`target_info` and `build_info` as follows:
`{__name__=~"(target|build)_info"}`. However, the identifying labels always
have to be `instance` and `job`.
These limitations are partially defeating the purpose of the `info` function.
At the current stage, this is an experiment to find out how useful the approach
turns out to be in practice. A final version of the `info` function will indeed
consider all matching info series and with their appropriate identifying labels.
## `irate()`
`irate(v range-vector)` calculates the per-second instant rate of increase of
the time series in the range vector. This is based on the last two data points.
Breaks in monotonicity (such as counter resets due to target restarts) are
automatically adjusted for.
The following example expression returns the per-second rate of HTTP requests
looking up to 5 minutes back for the two most recent data points, per time
series in the range vector:
```
irate(http_requests_total{job="api-server"}[5m])
```
`irate` should only be used when graphing volatile, fast-moving counters.
Use `rate` for alerts and slow-moving counters, as brief changes
in the rate can reset the `FOR` clause and graphs consisting entirely of rare
spikes are hard to read.
Note that when combining `irate()` with an
[aggregation operator](operators.md#aggregation-operators) (e.g. `sum()`)
or a function aggregating over time (any function ending in `_over_time`),
always take a `irate()` first, then aggregate. Otherwise `irate()` cannot detect
counter resets when your target restarts.
## `label_join()`
For each timeseries in `v`, `label_join(v instant-vector, dst_label string, separator string, src_label_1 string, src_label_2 string, ...)` joins all the values of all the `src_labels`
using `separator` and returns the timeseries with the label `dst_label` containing the joined value.
There can be any number of `src_labels` in this function.
promql: Separate `Point` into `FPoint` and `HPoint` In other words: Instead of having a “polymorphous” `Point` that can either contain a float value or a histogram value, use an `FPoint` for floats and an `HPoint` for histograms. This seemingly small change has a _lot_ of repercussions throughout the codebase. The idea here is to avoid the increase in size of `Point` arrays that happened after native histograms had been added. The higher-level data structures (`Sample`, `Series`, etc.) are still “polymorphous”. The same idea could be applied to them, but at each step the trade-offs needed to be evaluated. The idea with this change is to do the minimum necessary to get back to pre-histogram performance for functions that do not touch histograms. Here are comparisons for the `changes` function. The test data doesn't include histograms yet. Ideally, there would be no change in the benchmark result at all. First runtime v2.39 compared to directly prior to this commit: ``` name old time/op new time/op delta RangeQuery/expr=changes(a_one[1d]),steps=1-16 391µs ± 2% 542µs ± 1% +38.58% (p=0.000 n=9+8) RangeQuery/expr=changes(a_one[1d]),steps=10-16 452µs ± 2% 617µs ± 2% +36.48% (p=0.000 n=10+10) RangeQuery/expr=changes(a_one[1d]),steps=100-16 1.12ms ± 1% 1.36ms ± 2% +21.58% (p=0.000 n=8+10) RangeQuery/expr=changes(a_one[1d]),steps=1000-16 7.83ms ± 1% 8.94ms ± 1% +14.21% (p=0.000 n=10+10) RangeQuery/expr=changes(a_ten[1d]),steps=1-16 2.98ms ± 0% 3.30ms ± 1% +10.67% (p=0.000 n=9+10) RangeQuery/expr=changes(a_ten[1d]),steps=10-16 3.66ms ± 1% 4.10ms ± 1% +11.82% (p=0.000 n=10+10) RangeQuery/expr=changes(a_ten[1d]),steps=100-16 10.5ms ± 0% 11.8ms ± 1% +12.50% (p=0.000 n=8+10) RangeQuery/expr=changes(a_ten[1d]),steps=1000-16 77.6ms ± 1% 87.4ms ± 1% +12.63% (p=0.000 n=9+9) RangeQuery/expr=changes(a_hundred[1d]),steps=1-16 30.4ms ± 2% 32.8ms ± 1% +8.01% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=10-16 37.1ms ± 2% 40.6ms ± 2% +9.64% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=100-16 105ms ± 1% 117ms ± 1% +11.69% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=1000-16 783ms ± 3% 876ms ± 1% +11.83% (p=0.000 n=9+10) ``` And then runtime v2.39 compared to after this commit: ``` name old time/op new time/op delta RangeQuery/expr=changes(a_one[1d]),steps=1-16 391µs ± 2% 547µs ± 1% +39.84% (p=0.000 n=9+8) RangeQuery/expr=changes(a_one[1d]),steps=10-16 452µs ± 2% 616µs ± 2% +36.15% (p=0.000 n=10+10) RangeQuery/expr=changes(a_one[1d]),steps=100-16 1.12ms ± 1% 1.26ms ± 1% +12.20% (p=0.000 n=8+10) RangeQuery/expr=changes(a_one[1d]),steps=1000-16 7.83ms ± 1% 7.95ms ± 1% +1.59% (p=0.000 n=10+8) RangeQuery/expr=changes(a_ten[1d]),steps=1-16 2.98ms ± 0% 3.38ms ± 2% +13.49% (p=0.000 n=9+10) RangeQuery/expr=changes(a_ten[1d]),steps=10-16 3.66ms ± 1% 4.02ms ± 1% +9.80% (p=0.000 n=10+9) RangeQuery/expr=changes(a_ten[1d]),steps=100-16 10.5ms ± 0% 10.8ms ± 1% +3.08% (p=0.000 n=8+10) RangeQuery/expr=changes(a_ten[1d]),steps=1000-16 77.6ms ± 1% 78.1ms ± 1% +0.58% (p=0.035 n=9+10) RangeQuery/expr=changes(a_hundred[1d]),steps=1-16 30.4ms ± 2% 33.5ms ± 4% +10.18% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=10-16 37.1ms ± 2% 40.0ms ± 1% +7.98% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=100-16 105ms ± 1% 107ms ± 1% +1.92% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=1000-16 783ms ± 3% 775ms ± 1% -1.02% (p=0.019 n=9+9) ``` In summary, the runtime doesn't really improve with this change for queries with just a few steps. For queries with many steps, this commit essentially reinstates the old performance. This is good because the many-step queries are the one that matter most (longest absolute runtime). In terms of allocations, though, this commit doesn't make a dent at all (numbers not shown). The reason is that most of the allocations happen in the sampleRingIterator (in the storage package), which has to be addressed in a separate commit. Signed-off-by: beorn7 <beorn@grafana.com>
2022-10-28 14:58:40 +00:00
`label_join` acts on float and histogram samples in the same way.
This example will return a vector with each time series having a `foo` label with the value `a,b,c` added to it:
```
label_join(up{job="api-server",src1="a",src2="b",src3="c"}, "foo", ",", "src1", "src2", "src3")
```
## `label_replace()`
For each timeseries in `v`, `label_replace(v instant-vector, dst_label string, replacement string, src_label string, regex string)`
matches the [regular expression](https://github.com/google/re2/wiki/Syntax) `regex` against the value of the label `src_label`. If it
matches, the value of the label `dst_label` in the returned timeseries will be the expansion
of `replacement`, together with the original labels in the input. Capturing groups in the
regular expression can be referenced with `$1`, `$2`, etc. Named capturing groups in the regular expression can be referenced with `$name` (where `name` is the capturing group name). If the regular expression doesn't match then the timeseries is returned unchanged.
promql: Separate `Point` into `FPoint` and `HPoint` In other words: Instead of having a “polymorphous” `Point` that can either contain a float value or a histogram value, use an `FPoint` for floats and an `HPoint` for histograms. This seemingly small change has a _lot_ of repercussions throughout the codebase. The idea here is to avoid the increase in size of `Point` arrays that happened after native histograms had been added. The higher-level data structures (`Sample`, `Series`, etc.) are still “polymorphous”. The same idea could be applied to them, but at each step the trade-offs needed to be evaluated. The idea with this change is to do the minimum necessary to get back to pre-histogram performance for functions that do not touch histograms. Here are comparisons for the `changes` function. The test data doesn't include histograms yet. Ideally, there would be no change in the benchmark result at all. First runtime v2.39 compared to directly prior to this commit: ``` name old time/op new time/op delta RangeQuery/expr=changes(a_one[1d]),steps=1-16 391µs ± 2% 542µs ± 1% +38.58% (p=0.000 n=9+8) RangeQuery/expr=changes(a_one[1d]),steps=10-16 452µs ± 2% 617µs ± 2% +36.48% (p=0.000 n=10+10) RangeQuery/expr=changes(a_one[1d]),steps=100-16 1.12ms ± 1% 1.36ms ± 2% +21.58% (p=0.000 n=8+10) RangeQuery/expr=changes(a_one[1d]),steps=1000-16 7.83ms ± 1% 8.94ms ± 1% +14.21% (p=0.000 n=10+10) RangeQuery/expr=changes(a_ten[1d]),steps=1-16 2.98ms ± 0% 3.30ms ± 1% +10.67% (p=0.000 n=9+10) RangeQuery/expr=changes(a_ten[1d]),steps=10-16 3.66ms ± 1% 4.10ms ± 1% +11.82% (p=0.000 n=10+10) RangeQuery/expr=changes(a_ten[1d]),steps=100-16 10.5ms ± 0% 11.8ms ± 1% +12.50% (p=0.000 n=8+10) RangeQuery/expr=changes(a_ten[1d]),steps=1000-16 77.6ms ± 1% 87.4ms ± 1% +12.63% (p=0.000 n=9+9) RangeQuery/expr=changes(a_hundred[1d]),steps=1-16 30.4ms ± 2% 32.8ms ± 1% +8.01% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=10-16 37.1ms ± 2% 40.6ms ± 2% +9.64% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=100-16 105ms ± 1% 117ms ± 1% +11.69% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=1000-16 783ms ± 3% 876ms ± 1% +11.83% (p=0.000 n=9+10) ``` And then runtime v2.39 compared to after this commit: ``` name old time/op new time/op delta RangeQuery/expr=changes(a_one[1d]),steps=1-16 391µs ± 2% 547µs ± 1% +39.84% (p=0.000 n=9+8) RangeQuery/expr=changes(a_one[1d]),steps=10-16 452µs ± 2% 616µs ± 2% +36.15% (p=0.000 n=10+10) RangeQuery/expr=changes(a_one[1d]),steps=100-16 1.12ms ± 1% 1.26ms ± 1% +12.20% (p=0.000 n=8+10) RangeQuery/expr=changes(a_one[1d]),steps=1000-16 7.83ms ± 1% 7.95ms ± 1% +1.59% (p=0.000 n=10+8) RangeQuery/expr=changes(a_ten[1d]),steps=1-16 2.98ms ± 0% 3.38ms ± 2% +13.49% (p=0.000 n=9+10) RangeQuery/expr=changes(a_ten[1d]),steps=10-16 3.66ms ± 1% 4.02ms ± 1% +9.80% (p=0.000 n=10+9) RangeQuery/expr=changes(a_ten[1d]),steps=100-16 10.5ms ± 0% 10.8ms ± 1% +3.08% (p=0.000 n=8+10) RangeQuery/expr=changes(a_ten[1d]),steps=1000-16 77.6ms ± 1% 78.1ms ± 1% +0.58% (p=0.035 n=9+10) RangeQuery/expr=changes(a_hundred[1d]),steps=1-16 30.4ms ± 2% 33.5ms ± 4% +10.18% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=10-16 37.1ms ± 2% 40.0ms ± 1% +7.98% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=100-16 105ms ± 1% 107ms ± 1% +1.92% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=1000-16 783ms ± 3% 775ms ± 1% -1.02% (p=0.019 n=9+9) ``` In summary, the runtime doesn't really improve with this change for queries with just a few steps. For queries with many steps, this commit essentially reinstates the old performance. This is good because the many-step queries are the one that matter most (longest absolute runtime). In terms of allocations, though, this commit doesn't make a dent at all (numbers not shown). The reason is that most of the allocations happen in the sampleRingIterator (in the storage package), which has to be addressed in a separate commit. Signed-off-by: beorn7 <beorn@grafana.com>
2022-10-28 14:58:40 +00:00
`label_replace` acts on float and histogram samples in the same way.
This example will return timeseries with the values `a:c` at label `service` and `a` at label `foo`:
```
label_replace(up{job="api-server",service="a:c"}, "foo", "$1", "service", "(.*):.*")
```
This second example has the same effect than the first example, and illustrates use of named capturing groups:
```
label_replace(up{job="api-server",service="a:c"}, "foo", "$name", "service", "(?P<name>.*):(?P<version>.*)")
```
## `ln()`
`ln(v instant-vector)` calculates the natural logarithm for all elements in `v`.
Special cases are:
* `ln(+Inf) = +Inf`
* `ln(0) = -Inf`
* `ln(x < 0) = NaN`
* `ln(NaN) = NaN`
## `log2()`
`log2(v instant-vector)` calculates the binary logarithm for all elements in `v`.
The special cases are equivalent to those in `ln`.
## `log10()`
`log10(v instant-vector)` calculates the decimal logarithm for all elements in `v`.
The special cases are equivalent to those in `ln`.
## `minute()`
`minute(v=vector(time()) instant-vector)` returns the minute of the hour for each
of the given times in UTC. Returned values are from 0 to 59.
## `month()`
`month(v=vector(time()) instant-vector)` returns the month of the year for each
of the given times in UTC. Returned values are from 1 to 12, where 1 means
January etc.
## `predict_linear()`
`predict_linear(v range-vector, t scalar)` predicts the value of time series
`t` seconds from now, based on the range vector `v`, using [simple linear
regression](https://en.wikipedia.org/wiki/Simple_linear_regression).
The range vector must have at least two samples in order to perform the
calculation. When `+Inf` or `-Inf` are found in the range vector,
the slope and offset value calculated will be `NaN`.
`predict_linear` should only be used with gauges.
## `rate()`
`rate(v range-vector)` calculates the per-second average rate of increase of the
time series in the range vector. Breaks in monotonicity (such as counter
resets due to target restarts) are automatically adjusted for. Also, the
calculation extrapolates to the ends of the time range, allowing for missed
scrapes or imperfect alignment of scrape cycles with the range's time period.
The following example expression returns the per-second rate of HTTP requests as measured
over the last 5 minutes, per time series in the range vector:
```
rate(http_requests_total{job="api-server"}[5m])
```
`rate` acts on native histograms by calculating a new histogram where each
component (sum and count of observations, buckets) is the rate of increase
between the respective component in the first and last native histogram in
`v`. However, each element in `v` that contains a mix of float and native
histogram samples within the range, will be missing from the result vector.
`rate` should only be used with counters and native histograms where the
components behave like counters. It is best suited for alerting, and for
graphing of slow-moving counters.
Note that when combining `rate()` with an aggregation operator (e.g. `sum()`)
or a function aggregating over time (any function ending in `_over_time`),
always take a `rate()` first, then aggregate. Otherwise `rate()` cannot detect
counter resets when your target restarts.
## `resets()`
For each input time series, `resets(v range-vector)` returns the number of
counter resets within the provided time range as an instant vector. Any
promql: Separate `Point` into `FPoint` and `HPoint` In other words: Instead of having a “polymorphous” `Point` that can either contain a float value or a histogram value, use an `FPoint` for floats and an `HPoint` for histograms. This seemingly small change has a _lot_ of repercussions throughout the codebase. The idea here is to avoid the increase in size of `Point` arrays that happened after native histograms had been added. The higher-level data structures (`Sample`, `Series`, etc.) are still “polymorphous”. The same idea could be applied to them, but at each step the trade-offs needed to be evaluated. The idea with this change is to do the minimum necessary to get back to pre-histogram performance for functions that do not touch histograms. Here are comparisons for the `changes` function. The test data doesn't include histograms yet. Ideally, there would be no change in the benchmark result at all. First runtime v2.39 compared to directly prior to this commit: ``` name old time/op new time/op delta RangeQuery/expr=changes(a_one[1d]),steps=1-16 391µs ± 2% 542µs ± 1% +38.58% (p=0.000 n=9+8) RangeQuery/expr=changes(a_one[1d]),steps=10-16 452µs ± 2% 617µs ± 2% +36.48% (p=0.000 n=10+10) RangeQuery/expr=changes(a_one[1d]),steps=100-16 1.12ms ± 1% 1.36ms ± 2% +21.58% (p=0.000 n=8+10) RangeQuery/expr=changes(a_one[1d]),steps=1000-16 7.83ms ± 1% 8.94ms ± 1% +14.21% (p=0.000 n=10+10) RangeQuery/expr=changes(a_ten[1d]),steps=1-16 2.98ms ± 0% 3.30ms ± 1% +10.67% (p=0.000 n=9+10) RangeQuery/expr=changes(a_ten[1d]),steps=10-16 3.66ms ± 1% 4.10ms ± 1% +11.82% (p=0.000 n=10+10) RangeQuery/expr=changes(a_ten[1d]),steps=100-16 10.5ms ± 0% 11.8ms ± 1% +12.50% (p=0.000 n=8+10) RangeQuery/expr=changes(a_ten[1d]),steps=1000-16 77.6ms ± 1% 87.4ms ± 1% +12.63% (p=0.000 n=9+9) RangeQuery/expr=changes(a_hundred[1d]),steps=1-16 30.4ms ± 2% 32.8ms ± 1% +8.01% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=10-16 37.1ms ± 2% 40.6ms ± 2% +9.64% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=100-16 105ms ± 1% 117ms ± 1% +11.69% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=1000-16 783ms ± 3% 876ms ± 1% +11.83% (p=0.000 n=9+10) ``` And then runtime v2.39 compared to after this commit: ``` name old time/op new time/op delta RangeQuery/expr=changes(a_one[1d]),steps=1-16 391µs ± 2% 547µs ± 1% +39.84% (p=0.000 n=9+8) RangeQuery/expr=changes(a_one[1d]),steps=10-16 452µs ± 2% 616µs ± 2% +36.15% (p=0.000 n=10+10) RangeQuery/expr=changes(a_one[1d]),steps=100-16 1.12ms ± 1% 1.26ms ± 1% +12.20% (p=0.000 n=8+10) RangeQuery/expr=changes(a_one[1d]),steps=1000-16 7.83ms ± 1% 7.95ms ± 1% +1.59% (p=0.000 n=10+8) RangeQuery/expr=changes(a_ten[1d]),steps=1-16 2.98ms ± 0% 3.38ms ± 2% +13.49% (p=0.000 n=9+10) RangeQuery/expr=changes(a_ten[1d]),steps=10-16 3.66ms ± 1% 4.02ms ± 1% +9.80% (p=0.000 n=10+9) RangeQuery/expr=changes(a_ten[1d]),steps=100-16 10.5ms ± 0% 10.8ms ± 1% +3.08% (p=0.000 n=8+10) RangeQuery/expr=changes(a_ten[1d]),steps=1000-16 77.6ms ± 1% 78.1ms ± 1% +0.58% (p=0.035 n=9+10) RangeQuery/expr=changes(a_hundred[1d]),steps=1-16 30.4ms ± 2% 33.5ms ± 4% +10.18% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=10-16 37.1ms ± 2% 40.0ms ± 1% +7.98% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=100-16 105ms ± 1% 107ms ± 1% +1.92% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=1000-16 783ms ± 3% 775ms ± 1% -1.02% (p=0.019 n=9+9) ``` In summary, the runtime doesn't really improve with this change for queries with just a few steps. For queries with many steps, this commit essentially reinstates the old performance. This is good because the many-step queries are the one that matter most (longest absolute runtime). In terms of allocations, though, this commit doesn't make a dent at all (numbers not shown). The reason is that most of the allocations happen in the sampleRingIterator (in the storage package), which has to be addressed in a separate commit. Signed-off-by: beorn7 <beorn@grafana.com>
2022-10-28 14:58:40 +00:00
decrease in the value between two consecutive float samples is interpreted as a
counter reset. A reset in a native histogram is detected in a more complex way:
Any decrease in any bucket, including the zero bucket, or in the count of
observation constitutes a counter reset, but also the disappearance of any
previously populated bucket, an increase in bucket resolution, or a decrease of
the zero-bucket width.
`resets` should only be used with counters and counter-like native
histograms.
If the range vector contains a mix of float and histogram samples for the same
series, counter resets are detected separately and their numbers added up. The
change from a float to a histogram sample is _not_ considered a counter
reset. Each float sample is compared to the next float sample, and each
histogram is comprared to the next histogram.
## `round()`
`round(v instant-vector, to_nearest=1 scalar)` rounds the sample values of all
elements in `v` to the nearest integer. Ties are resolved by rounding up. The
optional `to_nearest` argument allows specifying the nearest multiple to which
the sample values should be rounded. This multiple may also be a fraction.
## `scalar()`
Given a single-element input vector, `scalar(v instant-vector)` returns the
sample value of that single element as a scalar. If the input vector does not
have exactly one element, `scalar` will return `NaN`.
## `sgn()`
`sgn(v instant-vector)` returns a vector with all sample values converted to their sign, defined as this: 1 if v is positive, -1 if v is negative and 0 if v is equal to zero.
## `sort()`
`sort(v instant-vector)` returns vector elements sorted by their sample values,
promql: Separate `Point` into `FPoint` and `HPoint` In other words: Instead of having a “polymorphous” `Point` that can either contain a float value or a histogram value, use an `FPoint` for floats and an `HPoint` for histograms. This seemingly small change has a _lot_ of repercussions throughout the codebase. The idea here is to avoid the increase in size of `Point` arrays that happened after native histograms had been added. The higher-level data structures (`Sample`, `Series`, etc.) are still “polymorphous”. The same idea could be applied to them, but at each step the trade-offs needed to be evaluated. The idea with this change is to do the minimum necessary to get back to pre-histogram performance for functions that do not touch histograms. Here are comparisons for the `changes` function. The test data doesn't include histograms yet. Ideally, there would be no change in the benchmark result at all. First runtime v2.39 compared to directly prior to this commit: ``` name old time/op new time/op delta RangeQuery/expr=changes(a_one[1d]),steps=1-16 391µs ± 2% 542µs ± 1% +38.58% (p=0.000 n=9+8) RangeQuery/expr=changes(a_one[1d]),steps=10-16 452µs ± 2% 617µs ± 2% +36.48% (p=0.000 n=10+10) RangeQuery/expr=changes(a_one[1d]),steps=100-16 1.12ms ± 1% 1.36ms ± 2% +21.58% (p=0.000 n=8+10) RangeQuery/expr=changes(a_one[1d]),steps=1000-16 7.83ms ± 1% 8.94ms ± 1% +14.21% (p=0.000 n=10+10) RangeQuery/expr=changes(a_ten[1d]),steps=1-16 2.98ms ± 0% 3.30ms ± 1% +10.67% (p=0.000 n=9+10) RangeQuery/expr=changes(a_ten[1d]),steps=10-16 3.66ms ± 1% 4.10ms ± 1% +11.82% (p=0.000 n=10+10) RangeQuery/expr=changes(a_ten[1d]),steps=100-16 10.5ms ± 0% 11.8ms ± 1% +12.50% (p=0.000 n=8+10) RangeQuery/expr=changes(a_ten[1d]),steps=1000-16 77.6ms ± 1% 87.4ms ± 1% +12.63% (p=0.000 n=9+9) RangeQuery/expr=changes(a_hundred[1d]),steps=1-16 30.4ms ± 2% 32.8ms ± 1% +8.01% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=10-16 37.1ms ± 2% 40.6ms ± 2% +9.64% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=100-16 105ms ± 1% 117ms ± 1% +11.69% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=1000-16 783ms ± 3% 876ms ± 1% +11.83% (p=0.000 n=9+10) ``` And then runtime v2.39 compared to after this commit: ``` name old time/op new time/op delta RangeQuery/expr=changes(a_one[1d]),steps=1-16 391µs ± 2% 547µs ± 1% +39.84% (p=0.000 n=9+8) RangeQuery/expr=changes(a_one[1d]),steps=10-16 452µs ± 2% 616µs ± 2% +36.15% (p=0.000 n=10+10) RangeQuery/expr=changes(a_one[1d]),steps=100-16 1.12ms ± 1% 1.26ms ± 1% +12.20% (p=0.000 n=8+10) RangeQuery/expr=changes(a_one[1d]),steps=1000-16 7.83ms ± 1% 7.95ms ± 1% +1.59% (p=0.000 n=10+8) RangeQuery/expr=changes(a_ten[1d]),steps=1-16 2.98ms ± 0% 3.38ms ± 2% +13.49% (p=0.000 n=9+10) RangeQuery/expr=changes(a_ten[1d]),steps=10-16 3.66ms ± 1% 4.02ms ± 1% +9.80% (p=0.000 n=10+9) RangeQuery/expr=changes(a_ten[1d]),steps=100-16 10.5ms ± 0% 10.8ms ± 1% +3.08% (p=0.000 n=8+10) RangeQuery/expr=changes(a_ten[1d]),steps=1000-16 77.6ms ± 1% 78.1ms ± 1% +0.58% (p=0.035 n=9+10) RangeQuery/expr=changes(a_hundred[1d]),steps=1-16 30.4ms ± 2% 33.5ms ± 4% +10.18% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=10-16 37.1ms ± 2% 40.0ms ± 1% +7.98% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=100-16 105ms ± 1% 107ms ± 1% +1.92% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=1000-16 783ms ± 3% 775ms ± 1% -1.02% (p=0.019 n=9+9) ``` In summary, the runtime doesn't really improve with this change for queries with just a few steps. For queries with many steps, this commit essentially reinstates the old performance. This is good because the many-step queries are the one that matter most (longest absolute runtime). In terms of allocations, though, this commit doesn't make a dent at all (numbers not shown). The reason is that most of the allocations happen in the sampleRingIterator (in the storage package), which has to be addressed in a separate commit. Signed-off-by: beorn7 <beorn@grafana.com>
2022-10-28 14:58:40 +00:00
in ascending order. Native histograms are sorted by their sum of observations.
Please note that `sort` only affects the results of instant queries, as range query results always have a fixed output ordering.
## `sort_desc()`
Same as `sort`, but sorts in descending order.
Like `sort`, `sort_desc` only affects the results of instant queries, as range query results always have a fixed output ordering.
## `sort_by_label()`
**This function has to be enabled via the [feature flag](../feature_flags.md#experimental-promql-functions) `--enable-feature=promql-experimental-functions`.**
`sort_by_label(v instant-vector, label string, ...)` returns vector elements sorted by the values of the given labels in ascending order. In case these label values are equal, elements are sorted by their full label sets.
Please note that the sort by label functions only affect the results of instant queries, as range query results always have a fixed output ordering.
This function uses [natural sort order](https://en.wikipedia.org/wiki/Natural_sort_order).
## `sort_by_label_desc()`
**This function has to be enabled via the [feature flag](../feature_flags.md#experimental-promql-functions) `--enable-feature=promql-experimental-functions`.**
Same as `sort_by_label`, but sorts in descending order.
Please note that the sort by label functions only affect the results of instant queries, as range query results always have a fixed output ordering.
This function uses [natural sort order](https://en.wikipedia.org/wiki/Natural_sort_order).
## `sqrt()`
`sqrt(v instant-vector)` calculates the square root of all elements in `v`.
## `time()`
`time()` returns the number of seconds since January 1, 1970 UTC. Note that
this does not actually return the current time, but the time at which the
expression is to be evaluated.
2017-10-27 14:44:57 +00:00
## `timestamp()`
`timestamp(v instant-vector)` returns the timestamp of each of the samples of
promql: Separate `Point` into `FPoint` and `HPoint` In other words: Instead of having a “polymorphous” `Point` that can either contain a float value or a histogram value, use an `FPoint` for floats and an `HPoint` for histograms. This seemingly small change has a _lot_ of repercussions throughout the codebase. The idea here is to avoid the increase in size of `Point` arrays that happened after native histograms had been added. The higher-level data structures (`Sample`, `Series`, etc.) are still “polymorphous”. The same idea could be applied to them, but at each step the trade-offs needed to be evaluated. The idea with this change is to do the minimum necessary to get back to pre-histogram performance for functions that do not touch histograms. Here are comparisons for the `changes` function. The test data doesn't include histograms yet. Ideally, there would be no change in the benchmark result at all. First runtime v2.39 compared to directly prior to this commit: ``` name old time/op new time/op delta RangeQuery/expr=changes(a_one[1d]),steps=1-16 391µs ± 2% 542µs ± 1% +38.58% (p=0.000 n=9+8) RangeQuery/expr=changes(a_one[1d]),steps=10-16 452µs ± 2% 617µs ± 2% +36.48% (p=0.000 n=10+10) RangeQuery/expr=changes(a_one[1d]),steps=100-16 1.12ms ± 1% 1.36ms ± 2% +21.58% (p=0.000 n=8+10) RangeQuery/expr=changes(a_one[1d]),steps=1000-16 7.83ms ± 1% 8.94ms ± 1% +14.21% (p=0.000 n=10+10) RangeQuery/expr=changes(a_ten[1d]),steps=1-16 2.98ms ± 0% 3.30ms ± 1% +10.67% (p=0.000 n=9+10) RangeQuery/expr=changes(a_ten[1d]),steps=10-16 3.66ms ± 1% 4.10ms ± 1% +11.82% (p=0.000 n=10+10) RangeQuery/expr=changes(a_ten[1d]),steps=100-16 10.5ms ± 0% 11.8ms ± 1% +12.50% (p=0.000 n=8+10) RangeQuery/expr=changes(a_ten[1d]),steps=1000-16 77.6ms ± 1% 87.4ms ± 1% +12.63% (p=0.000 n=9+9) RangeQuery/expr=changes(a_hundred[1d]),steps=1-16 30.4ms ± 2% 32.8ms ± 1% +8.01% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=10-16 37.1ms ± 2% 40.6ms ± 2% +9.64% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=100-16 105ms ± 1% 117ms ± 1% +11.69% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=1000-16 783ms ± 3% 876ms ± 1% +11.83% (p=0.000 n=9+10) ``` And then runtime v2.39 compared to after this commit: ``` name old time/op new time/op delta RangeQuery/expr=changes(a_one[1d]),steps=1-16 391µs ± 2% 547µs ± 1% +39.84% (p=0.000 n=9+8) RangeQuery/expr=changes(a_one[1d]),steps=10-16 452µs ± 2% 616µs ± 2% +36.15% (p=0.000 n=10+10) RangeQuery/expr=changes(a_one[1d]),steps=100-16 1.12ms ± 1% 1.26ms ± 1% +12.20% (p=0.000 n=8+10) RangeQuery/expr=changes(a_one[1d]),steps=1000-16 7.83ms ± 1% 7.95ms ± 1% +1.59% (p=0.000 n=10+8) RangeQuery/expr=changes(a_ten[1d]),steps=1-16 2.98ms ± 0% 3.38ms ± 2% +13.49% (p=0.000 n=9+10) RangeQuery/expr=changes(a_ten[1d]),steps=10-16 3.66ms ± 1% 4.02ms ± 1% +9.80% (p=0.000 n=10+9) RangeQuery/expr=changes(a_ten[1d]),steps=100-16 10.5ms ± 0% 10.8ms ± 1% +3.08% (p=0.000 n=8+10) RangeQuery/expr=changes(a_ten[1d]),steps=1000-16 77.6ms ± 1% 78.1ms ± 1% +0.58% (p=0.035 n=9+10) RangeQuery/expr=changes(a_hundred[1d]),steps=1-16 30.4ms ± 2% 33.5ms ± 4% +10.18% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=10-16 37.1ms ± 2% 40.0ms ± 1% +7.98% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=100-16 105ms ± 1% 107ms ± 1% +1.92% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=1000-16 783ms ± 3% 775ms ± 1% -1.02% (p=0.019 n=9+9) ``` In summary, the runtime doesn't really improve with this change for queries with just a few steps. For queries with many steps, this commit essentially reinstates the old performance. This is good because the many-step queries are the one that matter most (longest absolute runtime). In terms of allocations, though, this commit doesn't make a dent at all (numbers not shown). The reason is that most of the allocations happen in the sampleRingIterator (in the storage package), which has to be addressed in a separate commit. Signed-off-by: beorn7 <beorn@grafana.com>
2022-10-28 14:58:40 +00:00
the given vector as the number of seconds since January 1, 1970 UTC. It also
works with histogram samples.
2017-10-27 14:44:57 +00:00
## `vector()`
`vector(s scalar)` returns the scalar `s` as a vector with no labels.
## `year()`
`year(v=vector(time()) instant-vector)` returns the year
for each of the given times in UTC.
## `<aggregation>_over_time()`
The following functions allow aggregating each series of a given range vector
over time and return an instant vector with per-series aggregation results:
* `avg_over_time(range-vector)`: the average value of all points in the specified interval.
* `min_over_time(range-vector)`: the minimum value of all points in the specified interval.
* `max_over_time(range-vector)`: the maximum value of all points in the specified interval.
* `sum_over_time(range-vector)`: the sum of all values in the specified interval.
* `count_over_time(range-vector)`: the count of all values in the specified interval.
* `quantile_over_time(scalar, range-vector)`: the φ-quantile (0 ≤ φ ≤ 1) of the values in the specified interval.
* `stddev_over_time(range-vector)`: the population standard deviation of the values in the specified interval.
* `stdvar_over_time(range-vector)`: the population standard variance of the values in the specified interval.
promql: Separate `Point` into `FPoint` and `HPoint` In other words: Instead of having a “polymorphous” `Point` that can either contain a float value or a histogram value, use an `FPoint` for floats and an `HPoint` for histograms. This seemingly small change has a _lot_ of repercussions throughout the codebase. The idea here is to avoid the increase in size of `Point` arrays that happened after native histograms had been added. The higher-level data structures (`Sample`, `Series`, etc.) are still “polymorphous”. The same idea could be applied to them, but at each step the trade-offs needed to be evaluated. The idea with this change is to do the minimum necessary to get back to pre-histogram performance for functions that do not touch histograms. Here are comparisons for the `changes` function. The test data doesn't include histograms yet. Ideally, there would be no change in the benchmark result at all. First runtime v2.39 compared to directly prior to this commit: ``` name old time/op new time/op delta RangeQuery/expr=changes(a_one[1d]),steps=1-16 391µs ± 2% 542µs ± 1% +38.58% (p=0.000 n=9+8) RangeQuery/expr=changes(a_one[1d]),steps=10-16 452µs ± 2% 617µs ± 2% +36.48% (p=0.000 n=10+10) RangeQuery/expr=changes(a_one[1d]),steps=100-16 1.12ms ± 1% 1.36ms ± 2% +21.58% (p=0.000 n=8+10) RangeQuery/expr=changes(a_one[1d]),steps=1000-16 7.83ms ± 1% 8.94ms ± 1% +14.21% (p=0.000 n=10+10) RangeQuery/expr=changes(a_ten[1d]),steps=1-16 2.98ms ± 0% 3.30ms ± 1% +10.67% (p=0.000 n=9+10) RangeQuery/expr=changes(a_ten[1d]),steps=10-16 3.66ms ± 1% 4.10ms ± 1% +11.82% (p=0.000 n=10+10) RangeQuery/expr=changes(a_ten[1d]),steps=100-16 10.5ms ± 0% 11.8ms ± 1% +12.50% (p=0.000 n=8+10) RangeQuery/expr=changes(a_ten[1d]),steps=1000-16 77.6ms ± 1% 87.4ms ± 1% +12.63% (p=0.000 n=9+9) RangeQuery/expr=changes(a_hundred[1d]),steps=1-16 30.4ms ± 2% 32.8ms ± 1% +8.01% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=10-16 37.1ms ± 2% 40.6ms ± 2% +9.64% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=100-16 105ms ± 1% 117ms ± 1% +11.69% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=1000-16 783ms ± 3% 876ms ± 1% +11.83% (p=0.000 n=9+10) ``` And then runtime v2.39 compared to after this commit: ``` name old time/op new time/op delta RangeQuery/expr=changes(a_one[1d]),steps=1-16 391µs ± 2% 547µs ± 1% +39.84% (p=0.000 n=9+8) RangeQuery/expr=changes(a_one[1d]),steps=10-16 452µs ± 2% 616µs ± 2% +36.15% (p=0.000 n=10+10) RangeQuery/expr=changes(a_one[1d]),steps=100-16 1.12ms ± 1% 1.26ms ± 1% +12.20% (p=0.000 n=8+10) RangeQuery/expr=changes(a_one[1d]),steps=1000-16 7.83ms ± 1% 7.95ms ± 1% +1.59% (p=0.000 n=10+8) RangeQuery/expr=changes(a_ten[1d]),steps=1-16 2.98ms ± 0% 3.38ms ± 2% +13.49% (p=0.000 n=9+10) RangeQuery/expr=changes(a_ten[1d]),steps=10-16 3.66ms ± 1% 4.02ms ± 1% +9.80% (p=0.000 n=10+9) RangeQuery/expr=changes(a_ten[1d]),steps=100-16 10.5ms ± 0% 10.8ms ± 1% +3.08% (p=0.000 n=8+10) RangeQuery/expr=changes(a_ten[1d]),steps=1000-16 77.6ms ± 1% 78.1ms ± 1% +0.58% (p=0.035 n=9+10) RangeQuery/expr=changes(a_hundred[1d]),steps=1-16 30.4ms ± 2% 33.5ms ± 4% +10.18% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=10-16 37.1ms ± 2% 40.0ms ± 1% +7.98% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=100-16 105ms ± 1% 107ms ± 1% +1.92% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=1000-16 783ms ± 3% 775ms ± 1% -1.02% (p=0.019 n=9+9) ``` In summary, the runtime doesn't really improve with this change for queries with just a few steps. For queries with many steps, this commit essentially reinstates the old performance. This is good because the many-step queries are the one that matter most (longest absolute runtime). In terms of allocations, though, this commit doesn't make a dent at all (numbers not shown). The reason is that most of the allocations happen in the sampleRingIterator (in the storage package), which has to be addressed in a separate commit. Signed-off-by: beorn7 <beorn@grafana.com>
2022-10-28 14:58:40 +00:00
* `last_over_time(range-vector)`: the most recent point value in the specified interval.
* `present_over_time(range-vector)`: the value 1 for any series in the specified interval.
If the [feature flag](../feature_flags.md#experimental-promql-functions)
`--enable-feature=promql-experimental-functions` is set, the following
additional functions are available:
* `mad_over_time(range-vector)`: the median absolute deviation of all points in the specified interval.
Note that all values in the specified interval have the same weight in the
aggregation even if the values are not equally spaced throughout the interval.
`avg_over_time`, `sum_over_time`, `count_over_time`, `last_over_time`, and
`present_over_time` handle native histograms as expected. All other functions
ignore histogram samples.
promql: Separate `Point` into `FPoint` and `HPoint` In other words: Instead of having a “polymorphous” `Point` that can either contain a float value or a histogram value, use an `FPoint` for floats and an `HPoint` for histograms. This seemingly small change has a _lot_ of repercussions throughout the codebase. The idea here is to avoid the increase in size of `Point` arrays that happened after native histograms had been added. The higher-level data structures (`Sample`, `Series`, etc.) are still “polymorphous”. The same idea could be applied to them, but at each step the trade-offs needed to be evaluated. The idea with this change is to do the minimum necessary to get back to pre-histogram performance for functions that do not touch histograms. Here are comparisons for the `changes` function. The test data doesn't include histograms yet. Ideally, there would be no change in the benchmark result at all. First runtime v2.39 compared to directly prior to this commit: ``` name old time/op new time/op delta RangeQuery/expr=changes(a_one[1d]),steps=1-16 391µs ± 2% 542µs ± 1% +38.58% (p=0.000 n=9+8) RangeQuery/expr=changes(a_one[1d]),steps=10-16 452µs ± 2% 617µs ± 2% +36.48% (p=0.000 n=10+10) RangeQuery/expr=changes(a_one[1d]),steps=100-16 1.12ms ± 1% 1.36ms ± 2% +21.58% (p=0.000 n=8+10) RangeQuery/expr=changes(a_one[1d]),steps=1000-16 7.83ms ± 1% 8.94ms ± 1% +14.21% (p=0.000 n=10+10) RangeQuery/expr=changes(a_ten[1d]),steps=1-16 2.98ms ± 0% 3.30ms ± 1% +10.67% (p=0.000 n=9+10) RangeQuery/expr=changes(a_ten[1d]),steps=10-16 3.66ms ± 1% 4.10ms ± 1% +11.82% (p=0.000 n=10+10) RangeQuery/expr=changes(a_ten[1d]),steps=100-16 10.5ms ± 0% 11.8ms ± 1% +12.50% (p=0.000 n=8+10) RangeQuery/expr=changes(a_ten[1d]),steps=1000-16 77.6ms ± 1% 87.4ms ± 1% +12.63% (p=0.000 n=9+9) RangeQuery/expr=changes(a_hundred[1d]),steps=1-16 30.4ms ± 2% 32.8ms ± 1% +8.01% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=10-16 37.1ms ± 2% 40.6ms ± 2% +9.64% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=100-16 105ms ± 1% 117ms ± 1% +11.69% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=1000-16 783ms ± 3% 876ms ± 1% +11.83% (p=0.000 n=9+10) ``` And then runtime v2.39 compared to after this commit: ``` name old time/op new time/op delta RangeQuery/expr=changes(a_one[1d]),steps=1-16 391µs ± 2% 547µs ± 1% +39.84% (p=0.000 n=9+8) RangeQuery/expr=changes(a_one[1d]),steps=10-16 452µs ± 2% 616µs ± 2% +36.15% (p=0.000 n=10+10) RangeQuery/expr=changes(a_one[1d]),steps=100-16 1.12ms ± 1% 1.26ms ± 1% +12.20% (p=0.000 n=8+10) RangeQuery/expr=changes(a_one[1d]),steps=1000-16 7.83ms ± 1% 7.95ms ± 1% +1.59% (p=0.000 n=10+8) RangeQuery/expr=changes(a_ten[1d]),steps=1-16 2.98ms ± 0% 3.38ms ± 2% +13.49% (p=0.000 n=9+10) RangeQuery/expr=changes(a_ten[1d]),steps=10-16 3.66ms ± 1% 4.02ms ± 1% +9.80% (p=0.000 n=10+9) RangeQuery/expr=changes(a_ten[1d]),steps=100-16 10.5ms ± 0% 10.8ms ± 1% +3.08% (p=0.000 n=8+10) RangeQuery/expr=changes(a_ten[1d]),steps=1000-16 77.6ms ± 1% 78.1ms ± 1% +0.58% (p=0.035 n=9+10) RangeQuery/expr=changes(a_hundred[1d]),steps=1-16 30.4ms ± 2% 33.5ms ± 4% +10.18% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=10-16 37.1ms ± 2% 40.0ms ± 1% +7.98% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=100-16 105ms ± 1% 107ms ± 1% +1.92% (p=0.000 n=10+10) RangeQuery/expr=changes(a_hundred[1d]),steps=1000-16 783ms ± 3% 775ms ± 1% -1.02% (p=0.019 n=9+9) ``` In summary, the runtime doesn't really improve with this change for queries with just a few steps. For queries with many steps, this commit essentially reinstates the old performance. This is good because the many-step queries are the one that matter most (longest absolute runtime). In terms of allocations, though, this commit doesn't make a dent at all (numbers not shown). The reason is that most of the allocations happen in the sampleRingIterator (in the storage package), which has to be addressed in a separate commit. Signed-off-by: beorn7 <beorn@grafana.com>
2022-10-28 14:58:40 +00:00
## Trigonometric Functions
The trigonometric functions work in radians:
* `acos(v instant-vector)`: calculates the arccosine of all elements in `v` ([special cases](https://pkg.go.dev/math#Acos)).
* `acosh(v instant-vector)`: calculates the inverse hyperbolic cosine of all elements in `v` ([special cases](https://pkg.go.dev/math#Acosh)).
* `asin(v instant-vector)`: calculates the arcsine of all elements in `v` ([special cases](https://pkg.go.dev/math#Asin)).
* `asinh(v instant-vector)`: calculates the inverse hyperbolic sine of all elements in `v` ([special cases](https://pkg.go.dev/math#Asinh)).
* `atan(v instant-vector)`: calculates the arctangent of all elements in `v` ([special cases](https://pkg.go.dev/math#Atan)).
* `atanh(v instant-vector)`: calculates the inverse hyperbolic tangent of all elements in `v` ([special cases](https://pkg.go.dev/math#Atanh)).
* `cos(v instant-vector)`: calculates the cosine of all elements in `v` ([special cases](https://pkg.go.dev/math#Cos)).
* `cosh(v instant-vector)`: calculates the hyperbolic cosine of all elements in `v` ([special cases](https://pkg.go.dev/math#Cosh)).
* `sin(v instant-vector)`: calculates the sine of all elements in `v` ([special cases](https://pkg.go.dev/math#Sin)).
* `sinh(v instant-vector)`: calculates the hyperbolic sine of all elements in `v` ([special cases](https://pkg.go.dev/math#Sinh)).
* `tan(v instant-vector)`: calculates the tangent of all elements in `v` ([special cases](https://pkg.go.dev/math#Tan)).
* `tanh(v instant-vector)`: calculates the hyperbolic tangent of all elements in `v` ([special cases](https://pkg.go.dev/math#Tanh)).
The following are useful for converting between degrees and radians:
* `deg(v instant-vector)`: converts radians to degrees for all elements in `v`.
* `pi()`: returns pi.
* `rad(v instant-vector)`: converts degrees to radians for all elements in `v`.