It is possible that desired shards is always a bit higher than the
number of shards (less than 30%) and by exporting desired shards as the
raw number it will be easy to tell if a Prometheus is in that situation.
Signed-off-by: Chris Marchbanks <csmarchbanks@gmail.com>
* Refactor calculateDesiredShards + don't reshard if we're having issues
sending samples.
* Track lastSendTimestamp via an int64 with atomic add/load, add a test
for reshard calculation.
* Simplify conditional for skipping resharding, add samplesIn/Out to shard
testcase struct.
Signed-off-by: Callum Styan <callumstyan@gmail.com>
The WAL Watcher replays a checkpoint after it is created in order to
garbage collect series that no longer exist in the WAL. Currently the
garbage collection process is done serially with reading from the tip of
the WAL which can cause large delays in writing samples to remote
storage just after compaction occurs.
This also fixes a memory leak where dropped series are not cleaned up as
part of the SeriesReset process.
Signed-off-by: Chris Marchbanks <csmarchbanks@gmail.com>
* Replaced test validations with testutils on storage/remote/codec_test.go
Signed-off-by: George Felix <george.felix@ubeeqo.com>
* gofmt
Signed-off-by: George Felix <george.felix@ubeeqo.com>
* Removed shouldPass assertion
Signed-off-by: George Felix <gfelixc@gmail.com>
* Fixes to improve readability
Signed-off-by: George Felix <george.felix@ubeeqo.com>
* Fixes based on code review comments
Signed-off-by: George Felix <george.felix@ubeeqo.com>
* Added Fatal method and used it in buffer_test
Signed-off-by: Joe Elliott <number101010@gmail.com>
* Added period to meet contributing guidelines
Signed-off-by: Joe Elliott <number101010@gmail.com>
* Removed fatal testutil method. Refactored test cases to use testutil.Assert
Signed-off-by: Joe Elliott <number101010@gmail.com>
* Added if found condition for clarity
Signed-off-by: Joe Elliott <number101010@gmail.com>
* Show warnings in UI if query have returned some warnings
+ improve warning (error) text if query to remote was finished with error
* Add prefixes for remote_read errors
Signed-off-by: Stan Putrya <root.vagner@gmail.com>
* Update go.mod dependencies before release
Signed-off-by: Julius Volz <julius.volz@gmail.com>
* Add issue for showing query warnings in promtool
Signed-off-by: Julius Volz <julius.volz@gmail.com>
* Revert json-iterator back to 1.1.6
It produced errors when marshaling Point values with special float
values.
Signed-off-by: Julius Volz <julius.volz@gmail.com>
* Fix expected step values in promtool tests after client_golang update
Signed-off-by: Julius Volz <julius.volz@gmail.com>
* Update generated protobuf code after proto dep updates
Signed-off-by: Julius Volz <julius.volz@gmail.com>
The desired shards calculation now properly keeps track of the rate of
pending samples, and uses the previously unused integralAccumulator to
adjust for missing information in the desired shards calculation.
Also, configure more capacity for each shard. The default 10 capacity
causes shards to block on each other while
sending remote requests. Default to a 500 sample capacity and explain in
the documentation that having more capacity will help throughput.
Signed-off-by: Chris Marchbanks <csmarchbanks@gmail.com>
* Check for duplicate label names in remote read
Also add test to confirm that #5731 is fixed
* Use subtests in TestValidateLabelsAndMetricName
* Really check that expectedErr matches err
Signed-off-by: Vadym Martsynovskyy <vmartsynovskyy@gmail.com>
* Only check last directory when discovering checkpoint number
Signed-off-by: Devin Trejo <dtrejo@palantir.com>
* Comments for checkpointNum
Signed-off-by: Devin Trejo <dtrejo@palantir.com>
* Add benchmark for sample delivery
* Simplify StoreSeries to have only one loop
* Reduce allocations for pending samples in runShard
* Only allocate one send slice per segment
* Cache a buffer in each shard for snappy to use
* Remove queue manager seriesMtx
It is not possible for any of the places protected by the seriesMtx to
be called concurrently so it is safe to remove. By removing the mutex we
can simplify the Append code to one loop.
Signed-off-by: Chris Marchbanks <csmarchbanks@gmail.com>
This allows other processes to reuse just the remote write code without
having to use the remote read code as well. This will be used to create
a sidecar capable of sending remote write payloads.
Signed-off-by: Chris Marchbanks <csmarchbanks@gmail.com>
* Update remote write and remote read separately
* Add external labels to the remote write conf hash
* Add unit tests for remote storage lifecycle
Signed-off-by: Chris Marchbanks <csmarchbanks@gmail.com>
* Don't panic if we try to release a string that is not in the interner.
* Move seriesMtx locking in QueueManager's StoreSeries function.
This stops us from calling release for strings that aren't interned if
there's a race between reading a checkpoint and storing new series
labels, which could happen during checkpointing or reloading config.
Signed-off-by: Callum Styan <callumstyan@gmail.com>
* Unregister remote write queue manager specific metrics when stopping the
queue manager.
* Use DeleteLabelValues instead of Unregister to remove queue and watcher
related metrics when we stop them. Create those metrics in the structs
start functions rather than in their constructors because of the
ordering of creation, start, and stop in remote storage ApplyConfig.
* Add setMetrics function to WAL watcher so we can set
the watchers metrics in it's Start function, but not
have to call Start in some tests (causes data race).
Signed-off-by: Callum Styan <callumstyan@gmail.com>
From the documentation:
> The default HTTP client's Transport may not
> reuse HTTP/1.x "keep-alive" TCP connections if the Body is
> not read to completion and closed.
This effectively enable keep-alive for the fixed requests.
Signed-off-by: Romain Baugue <romain.baugue@elwinar.com>
* Fix ReadCheckpoint to ensure that it actually reads all the contents of
each segment in a checkpoint dir, or returns an error.
Signed-off-by: Callum Styan <callumstyan@gmail.com>
i) Uses the more idiomatic Wrap and Wrapf methods for creating nested errors.
ii) Fixes some incorrect usages of fmt.Errorf where the error messages don't have any formatting directives.
iii) Does away with the use of fmt package for errors in favour of pkg/errors
Signed-off-by: tariqibrahim <tariq181290@gmail.com>
a string when there are no longer any refs. Add tests for interning.
Co-authored-by: Tom Wilkie <tom.wilkie@gmail.com>
Signed-off-by: Callum Styan <callumstyan@gmail.com>
- Unmarshall external_labels config as labels.Labels, add tests.
- Convert some more uses of model.LabelSet to labels.Labels.
- Remove old relabel pkg (fixes#3647).
- Validate external label names.
Signed-off-by: Tom Wilkie <tom.wilkie@gmail.com>
* Consistently pre-lookup the metrics for a given queue in queue manager.
* Don't open the WAL (for writing) in the remote_write code.
* Add some more logging.
Signed-off-by: Tom Wilkie <tom.wilkie@gmail.com>
- Remove prometheus_remote_queue_last_send_timestamp_seconds metric. Its not particularly useful, we have highest_timestamp_seconds.
- Factor out maxGauage, a gauge that only increases.
- Change sharding calculations to use max samples in timestamp - max samples out timestamp (not rates).
- Also include the ratio of samples dropped to correctly predict number of pending samples.
Signed-off-by: Tom Wilkie <tom.wilkie@gmail.com>
- Add a dropped samples EWMA and use it in calculating desired shards.
- Update metric names and a log messages.
- Limit number of entries in the dedupe logging middleware to prevent potential OOM.
Signed-off-by: Callum Styan <callumstyan@gmail.com>
Signed-off-by: Tom Wilkie <tom.wilkie@gmail.com>
- If we're replaying the WAL to get series records, skip that segment when we hit corruptions.
- If we're tailing the WAL for samples, fail the watcher.
- When the watcher fails, restart from the latest checkpoint - and only send new samples by updating startTime.
- Tidy up log lines and error handling, don't return so many errors on quiting.
- Expect EOF when processing checkpoints.
Signed-off-by: Tom Wilkie <tom.wilkie@gmail.com>
- Tests that created a QueueManager were leaving behind files at the end of tests.
- WAL replaying (readToEnd)tests seem to require extra time to finish now.
- Some fixes to make staticcheck happy
Signed-off-by: Callum Styan <callumstyan@gmail.com>
* storage/remote: adapt tests for Travis CI
Signed-off-by: Simon Pasquier <spasquie@redhat.com>
* Check filesystems on Travis environment
Signed-off-by: Simon Pasquier <spasquie@redhat.com>
* Run remote/storage tests on CircleCI for troubleshooting
Signed-off-by: Simon Pasquier <spasquie@redhat.com>
* Try using tmpfs partition
Signed-off-by: Simon Pasquier <spasquie@redhat.com>
* Revert "Try using tmpfs partition"
This reverts commit 85a30deb72.
Signed-off-by: Simon Pasquier <spasquie@redhat.com>
* Don't store labels in writeToMock
Signed-off-by: Simon Pasquier <spasquie@redhat.com>
* Fix data race
Signed-off-by: Simon Pasquier <spasquie@redhat.com>
* Bump retries to 100 meaning that the total timeout is 10s
Signed-off-by: Simon Pasquier <spasquie@redhat.com>
* clean up .travis.yml
Signed-off-by: Simon Pasquier <spasquie@redhat.com>
* code fixup
Signed-off-by: Simon Pasquier <spasquie@redhat.com>
* Remove unneeded empty line
Signed-off-by: Simon Pasquier <spasquie@redhat.com>
- Use the queue name in WAL watcher logging.
- Don't return from watch if the reader error was EOF.
- Fix sample timestamp check logic regarding what samples we send.
- Refactor so we don't need readToEnd/readSeriesRecords
- Fix wal_watcher tests since readToEnd no longer exists
Signed-off-by: Callum Styan <callumstyan@gmail.com>
- Remove datarace in the exported highest scrape timestamp.
- Backoff on enqueue should be per-sample - reset the result for each sample.
- Remove diffKeys, unused ctx and cancelfunc in WALWatcher, 'name' from writeTo interface, and pass it to constructor.
- Reorder functions in WALWatcher depth-first according to call graph.
- Fix vendor/modules.txt.
- Split out the various timer periods into consts at the top of the file.
- Move w.currentSegmentMetric.Set close to where we set the currentSegment.
- Combine r.Next() and isClosed(w.quit) into a single loop.
- Unnest some ifs in WALWatcher.watch, propagate erros in decodeRecord, add some new lines to make it easier to read.
- Reorganise checkpoint handling to reduce nesting and make it easier to follow.
Signed-off-by: Tom Wilkie <tom.wilkie@gmail.com>
This change switches the remote_write API to use the TSDB WAL. This should reduce memory usage and prevent sample loss when the remote end point is down.
We use the new LiveReader from TSDB to tail WAL segments. Logic for finding the tracking segment is included in this PR. The WAL is tailed once for each remote_write endpoint specified. Reading from the segment is based on a ticker rather than relying on fsnotify write events, which were found to be complicated and unreliable in early prototypes.
Enqueuing a sample for sending via remote_write can now block, to provide back pressure. Queues are still required to acheive parallelism and batching. We have updated the queue config based on new defaults for queue capacity and pending samples values - much smaller values are now possible. The remote_write resharding code has been updated to prevent deadlocks, and extra tests have been added for these cases.
As part of this change, we attempt to guarantee that samples are not lost; however this initial version doesn't guarantee this across Prometheus restarts or non-retryable errors from the remote end (eg 400s).
This changes also includes the following optimisations:
- only marshal the proto request once, not once per retry
- maintain a single copy of the labels for given series to reduce GC pressure
Other minor tweaks:
- only reshard if we've also successfully sent recently
- add pending samples, latest sent timestamp, WAL events processed metrics
Co-authored-by: Chris Marchbanks <csmarchbanks.com> (initial prototype)
Co-authored-by: Tom Wilkie <tom.wilkie@gmail.com> (sharding changes)
Signed-off-by: Callum Styan <callumstyan@gmail.com>
* Add flag for size based retention
Signed-off-by: Goutham Veeramachaneni <gouthamve@gmail.com>
* Deprecate the old retention flag for a new one.
Signed-off-by: Goutham Veeramachaneni <gouthamve@gmail.com>
* Add ability to take a suffix for size flag
Signed-off-by: Goutham Veeramachaneni <gouthamve@gmail.com>
* Address feedback
Signed-off-by: Goutham Veeramachaneni <gouthamve@gmail.com>
* update tsdb to v0.4.0
Signed-off-by: Krasi Georgiev <kgeorgie@redhat.com>
* remove unused struct field
Signed-off-by: Krasi Georgiev <kgeorgie@redhat.com>
* *: use latest release of staticcheck
It also fixes a couple of things in the code flagged by the additional
checks.
Signed-off-by: Simon Pasquier <spasquie@redhat.com>
* Use official release of staticcheck
Also run 'go list' before staticcheck to avoid failures when downloading packages.
Signed-off-by: Simon Pasquier <spasquie@redhat.com>
Add WALSegmentSize as an option, and the corresponding flag "storage.tsdb.wal-segment-size" to tune the max size of wal segment files.
The addressed base problem is to reduce the disk space used by wal segment files : on a raspberry pi, for instance, we often want to reduce write load of the sd card, then, the wal directory is mounted on a memory (space limited) partition.
the default value of the segment max file size, pushed the size of directory to 128 MB for each segment , which is too much ram consumption on a rasp.
the initial discussion is at https://github.com/prometheus/tsdb/pull/450
Currently Prometheus requests show up with a UA of Go-http-client/1.1
which isn't super helpful. Though the X-Prometheus-Remote-* headers
exist they need to be explicitly configured when logging the request in
order to be able to deduce this is a request originating from
Prometheus. By setting the header we remove this ambiguity and make
default server logs just a bit more useful.
This also updates a few other places to consistently capitalize the 'P'
in the user agent, as well as ensure we set a UA to begin with.
Signed-off-by: Daniele Sluijters <daenney@users.noreply.github.com>
* *: remove use of golang.org/x/net/context
Signed-off-by: Simon Pasquier <spasquie@redhat.com>
* scrape: fix TestTargetScrapeScrapeCancel
Signed-off-by: Simon Pasquier <spasquie@redhat.com>
* Limit the number of samples remote read can return.
- Return 413 entity too large.
- Limit can be set be a flag. Allow 0 to mean no limit.
- Include limit in error message.
- Set default limit to 50M (* 16 bytes = 800MB).
Signed-off-by: Tom Wilkie <tom.wilkie@gmail.com>
There are many more (mostly finalizers like Close/Stop/etc.), but most of
the others seemed like one couldn't do much about them anyway.
Signed-off-by: Julius Volz <julius.volz@gmail.com>
* Allow for BufferedSeriesIterator instances to be created without an underlying iterator, to simplify their usage.
Signed-off-by: Alin Sinpalean <alin.sinpalean@gmail.com>
* Add Start/End to SelectParams
* Make remote read use the new selectParams for start/end
This commit will continue sending the start/end time of the remote read
query as the overarching promql time and the specific range of data that
the query is intersted in receiving a response to is now part of the
ReadHints (upstream discussion in #4226).
* Remove unused vendored code
The genproto.sh script was updated, but the code wasn't regenerated.
This simply removes the vendored deps that are no longer part of the
codegen output.
Signed-off-by: Thomas Jackson <jacksontj.89@gmail.com>
Fixes#4254
NaNs don't equal themselves, so a duplicate NaN would
always hit the break statement and never get popped.
We should not be returning multiple data point for the same
timestamp, so don't compare values at all.
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
This commit fixes a denial-of-service issue of the remote
read endpoint. It limits the size of the POST request body
to 32 MB such that clients cannot write arbitrary amounts
of data to the server memory.
Fixes#4238
Signed-off-by: Andreas Auernhammer <aead@mail.de>
* Move range logic to 'eval'
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Make aggregegate range aware
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* PromQL is statically typed, so don't eval to find the type.
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Extend rangewrapper to multiple exprs
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Start making function evaluation ranged
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Make instant queries a special case of range queries
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Eliminate evalString
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Evaluate range vector functions one series at a time
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Make unary operators range aware
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Make binops range aware
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Pass time to range-aware functions.
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Make simple _over_time functions range aware
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Reduce allocs when working with matrix selectors
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Add basic benchmark for range evaluation
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Reuse objects for function arguments
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Do dropmetricname and allocating output vector only once.
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Add range-aware support for range vector functions with params
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Optimise holt_winters, cut cpu and allocs by ~25%
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Make rate&friends range aware
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Make more functions range aware. Document calling convention.
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Make date functions range aware
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Make simple math functions range aware
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Convert more functions to be range aware
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Make more functions range aware
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Specialcase timestamp() with vector selector arg for range awareness
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Remove transition code for functions
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Remove the rest of the engine transition code
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Remove more obselete code
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Remove the last uses of the eval* functions
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Remove engine finalizers to prevent corruption
The finalizers set by matrixSelector were being called
just before the value they were retruning to the pool
was then being provided to the caller. Thus a concurrent query
could corrupt the data that the user has just been returned.
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Add new benchmark suite for range functinos
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Migrate existing benchmarks to new system
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Expand promql benchmarks
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Simply test by removing unused range code
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* When testing instant queries, check range queries too.
To protect against subsequent steps in a range query being
affected by the previous steps, add a test that evaluates
an instant query that we know works again as a range query
with the tiimestamp we care about not being the first step.
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Reuse ring for matrix iters. Put query results back in pool.
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Reuse buffer when iterating over matrix selectors
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Unary minus should remove metric name
Cut down benchmarks for faster runs.
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Reduce repetition in benchmark test cases
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Work series by series when doing normal vectorSelectors
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Optimise benchmark setup, cuts time by 60%
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Have rangeWrapper use an evalNodeHelper to cache across steps
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Use evalNodeHelper with functions
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Cache dropMetricName within a node evaluation.
This saves both the calculations and allocs done by dropMetricName
across steps.
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Reuse input vectors in rangewrapper
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Reuse the point slices in the matrixes input/output by rangeWrapper
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Make benchmark setup faster using AddFast
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Simplify benchmark code.
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Add caching in VectorBinop
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Use xor to have one-level resultMetric hash key
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Add more benchmarks
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Call Query.Close in apiv1
This allows point slices allocated for the response data
to be reused by later queries, saving allocations.
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Optimise histogram_quantile
It's now 5-10% faster with 97% less garbage generated for 1k steps
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Make the input collection in rangeVector linear rather than quadratic
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Optimise label_replace, for 1k steps 15x fewer allocs and 3x faster
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Optimise label_join, 1.8x faster and 11x less memory for 1k steps
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Expand benchmarks, cleanup comments, simplify numSteps logic.
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Address Fabian's comments
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Comments from Alin.
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Address jrv's comments
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Remove dead code
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Address Simon's comments.
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Rename populateIterators, pre-init some sizes
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Handle case where function has non-matrix args first
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Split rangeWrapper out to rangeEval function, improve comments
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Cleanup and make things more consistent
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Make EvalNodeHelper public
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
* Fabian's comments.
Signed-off-by: Brian Brazil <brian.brazil@robustperception.io>
More than one remote_write destination can be configured, in which
case it's essential to know which one each log message refers to.
Signed-off-by: Bryan Boreham <bjboreham@gmail.com>
This adds a parameter to the storage selection interface which allows
query engine(s) to pass information about the operations surrounding a
data selection.
This can for example be used by remote storage backends to infer the
correct downsampling aggregates that need to be provided.
* refactor: move targetGroup struct and CheckOverflow() to their own package
* refactor: move auth and security related structs to a utility package, fix import error in utility package
* refactor: Azure SD, remove SD struct from config
* refactor: DNS SD, remove SD struct from config into dns package
* refactor: ec2 SD, move SD struct from config into the ec2 package
* refactor: file SD, move SD struct from config to file discovery package
* refactor: gce, move SD struct from config to gce discovery package
* refactor: move HTTPClientConfig and URL into util/config, fix import error in httputil
* refactor: consul, move SD struct from config into consul discovery package
* refactor: marathon, move SD struct from config into marathon discovery package
* refactor: triton, move SD struct from config to triton discovery package, fix test
* refactor: zookeeper, move SD structs from config to zookeeper discovery package
* refactor: openstack, remove SD struct from config, move into openstack discovery package
* refactor: kubernetes, move SD struct from config into kubernetes discovery package
* refactor: notifier, use targetgroup package instead of config
* refactor: tests for file, marathon, triton SD - use targetgroup package instead of config.TargetGroup
* refactor: retrieval, use targetgroup package instead of config.TargetGroup
* refactor: storage, use config util package
* refactor: discovery manager, use targetgroup package instead of config.TargetGroup
* refactor: use HTTPClient and TLS config from configUtil instead of config
* refactor: tests, use targetgroup package instead of config.TargetGroup
* refactor: fix tagetgroup.Group pointers that were removed by mistake
* refactor: openstack, kubernetes: drop prefixes
* refactor: remove import aliases forced due to vscode bug
* refactor: move main SD struct out of config into discovery/config
* refactor: rename configUtil to config_util
* refactor: rename yamlUtil to yaml_config
* refactor: kubernetes, remove prefixes
* refactor: move the TargetGroup package to discovery/
* refactor: fix order of imports
Federation makes use of dedupedSeriesSet to merge SeriesSets for every
query into one output stream. If many match[] arguments are provided,
many dedupedSeriesSet objects will get chained. This has the downside of
causing a potential O(n*k) running time, where n is the number of series
and k the number of match[] arguments.
In the mean time, the storage package provides a mergeSeriesSet that
accomplishes the same with an O(n*log(k)) running time by making use of
a binary heap. Let's just get rid of dedupedSeriesSet and change all
existing callers to use mergeSeriesSet.
For special remote read endpoints which have only data for specific
queries, it is desired to limit the number of queries sent to the
configured remote read endpoint to reduce latency and performance
overhead.
* Decouple remote client from ReadRecent feature.
* Separate remote read filter into a small, testable function.
* Use storage.Queryable interface to compose independent
functionalities.
In order to compose different querier implementations more easily, this
change introduces a separate storage.Queryable interface grouping the
query (Querier) function of the storage.
Furthermore, it adds a QueryableFunc type to ease writing very simple
queryable implementations.
The labelsets returned from remote read are mutated in higher levels
(like seriesFilter.Labels()) and since the concreteSeriesSet didn't
return a copy, the external mutation affected the labelset in the
concreteSeries itself. This resulted in bizarre bugs where local and
remote series would show with identical label sets in the UI, but not be
deduplicated, since internally, a series might come to look like:
{__name__="node_load5", instance="192.168.1.202:12090", job="node_exporter", node="odroid", node="odroid"}
(note the repetition of the last label)
* Fast path the merge querier such that it is completely removed from query path when there is no remote storage.
* Add NoopQuerier
* Add copyright notice.
* Avoid global, use a function.
If the user accidentally sets the max block duration smaller than the min,
the current error is not informative. This change just performs the check
earlier and improves the error message.
staticcheck fails with:
storage/remote/read_test.go:199:27: do not pass a nil Context, even if a function permits it; pass context.TODO if you are unsure about which Context to use (SA1012)
Currently all read queries are simply pushed to remote read clients.
This is fine, except for remote storage for wich it unefficient and
make query slower even if remote read is unnecessary.
So we need instead to compare the oldest timestamp in primary/local
storage with the query range lower boundary. If the oldest timestamp
is older than the mint parameter, then there is no need for remote read.
This is an optionnal behavior per remote read client.
Signed-off-by: Thibault Chataigner <t.chataigner@criteo.com>
Instead, just make the anchoring part of the internal regex. This helps because
some users will want to read back the `Value` field and expect it to be the
same as the input value (e.g. some tests in Cortex), or use the value in
another context which is already expected to add its own anchoring, leading to
superfluous double anchoring (such as when we translate matchers into remote
read request matchers).
* Re-add contexts to storage.Storage.Querier()
These are needed when replacing the storage by a multi-tenant
implementation where the tenant is stored in the context.
The 1.x query interfaces already had contexts, but they got lost in 2.x.
* Convert promql.Engine to use native contexts
This can happen in the situation where the system scales up the number of shards massively (to deal with some backlog), then scales it down again as the number of samples sent during the time period is less than the number received.
* Fix error where we look into the future.
So currently we are adding values that are in the future for an older
timestamp. For example, if we have [(1, 1), (150, 2)] we will end up
showing [(1, 1), (2,2)].
Further it is not advisable to call .At() after Next() returns false.
Signed-off-by: Goutham Veeramachaneni <cs14btech11014@iith.ac.in>
* Retuen early if done
Signed-off-by: Goutham Veeramachaneni <cs14btech11014@iith.ac.in>
* Handle Seek() where we reach the end of iterator
Signed-off-by: Goutham Veeramachaneni <cs14btech11014@iith.ac.in>
* Simplify code
Signed-off-by: Goutham Veeramachaneni <cs14btech11014@iith.ac.in>
This is in line with the v1.5 change in paradigm to not keep
chunk.Descs without chunks around after a series maintenance.
It's mainly motivated by avoiding excessive amounts of RAM usage
during crash recovery.
The code avoids to create memory time series with zero chunk.Descs as
that is prone to trigger weird effects. (Series maintenance would
archive series with zero chunk.Descs, but we cannot do that here
because the archive indices still have to be checked.)
The fpIter was kind of cumbersome to use and required a lock for each
iteration (which wasn't even needed for the iteration at startup after
loading the checkpoint).
The new implementation here has an obvious penalty in memory, but it's
only 8 byte per series, so 80MiB for a beefy server with 10M memory
time series (which would probably need ~100GiB RAM, so the memory
penalty is only 0.1% of the total memory need).
The big advantage is that now series maintenance happens in order,
which leads to the time between two maintenances of the same series
being less random. Ideally, after each maintenance, the next
maintenance would tackle the series with the largest number of
non-persisted chunks. That would be quite an effort to find out or
track, but with the approach here, the next maintenance will tackle
the series whose previous maintenance is longest ago, which is a good
approximation.
While this commit won't change the _average_ number of chunks
persisted per maintenance, it will reduce the mean time a given chunk
has to wait for its persistence and thus reduce the steady-state
number of chunks waiting for persistence.
Also, the map iteration in Go is non-deterministic but not truly
random. In practice, the iteration appears to be somewhat "bucketed".
You can often observe a bunch of series with similar duration since
their last maintenance, i.e. you see batches of series with similar
number of chunks persisted per maintenance. If that batch is
relatively young, a whole lot of series are maintained with very few
chunks to persist. (See screenshot in PR for a better explanation.)
This is a fairly easy attempt to dynamically evict chunks based on the
heap size. A target heap size has to be set as a command line flage,
so that users can essentially say "utilize 4GiB of RAM, and please
don't OOM".
The -storage.local.max-chunks-to-persist and
-storage.local.memory-chunks flags are deprecated by this
change. Backwards compatibility is provided by ignoring
-storage.local.max-chunks-to-persist and use
-storage.local.memory-chunks to set the new
-storage.local.target-heap-size to a reasonable (and conservative)
value (both with a warning).
This also makes the metrics intstrumentation more consistent (in
naming and implementation) and cleans up a few quirks in the tests.
Answers to anticipated comments:
There is a chance that Go 1.9 will allow programs better control over
the Go memory management. I don't expect those changes to be in
contradiction with the approach here, but I do expect them to
complement them and allow them to be more precise and controlled. In
any case, once those Go changes are available, this code has to be
revisted.
One might be tempted to let the user specify an estimated value for
the RSS usage, and then internall set a target heap size of a certain
fraction of that. (In my experience, 2/3 is a fairly safe bet.)
However, investigations have shown that RSS size and its relation to
the heap size is really really complicated. It depends on so many
factors that I wouldn't even start listing them in a commit
description. It depends on many circumstances and not at least on the
risk trade-off of each individual user between RAM utilization and
probability of OOMing during a RAM usage peak. To not add even more to
the confusion, we need to stick to the well-defined number we also use
in the targeting here, the sum of the sizes of heap objects.
Currently, if a series stops to exist, its head chunk will be kept
open for an hour. That prevents it from being persisted. Which
prevents it from being evicted. Which prevents the series from being
archived.
Most of the time, once no sample has been added to a series within the
staleness limit, we can be pretty confident that this series will not
receive samples anymore. The whole chain as described above can be
started after 5m instead of 1h. In the relaxed case, this doesn't
change a lot as the head chunk timeout is only checked during series
maintenance, and usually, a series is only maintained every six
hours. However, there is the typical scenario where a large service is
deployed, the deoply turns out to be bad, and then it is deployed
again within minutes, and quite quickly the number of time series has
tripled. That's the point where the Prometheus server is stressed and
switches (rightfully) into rushed mode. In that mode, time series are
processed as quickly as possible, but all of that is in vein if all of
those recently ended time series cannot be persisted yet for another
hour. In that scenario, this change will help most, and it's exactly
the scenario where help is most desperately needed.
Each remote write endpoint gets its own set of relabeling rules.
This is based on the (yet-to-be-merged)
https://github.com/prometheus/prometheus/pull/2419, which removes legacy
remote write implementations.
This removes legacy support for specific remote storage systems in favor
of only offering the generic remote write protocol. An example bridge
application that translates from the generic protocol to each of those
legacy backends is still provided at:
documentation/examples/remote_storage/remote_storage_bridge
See also https://github.com/prometheus/prometheus/issues/10
The next step in the plan is to re-add support for multiple remote
storages.
This is another corner-case that was previously never exercised
because the rewriting of a series file was never prevented by the
shrink ratio.
Scenario: There is an existing series on disk, which is archived. If a
new sample comes in for that file, a new chunk in memory is created,
and the chunkDescsOffset is set to -1. If series maintenance happens
before the series has at least one chunk to persist _and_ an
insufficient chunks on disk is old enough for purging (so that the
shrink ratio kicks in), dropAndPersistChunks would return 0, but it
should return the chunk length of the series file.
Also, in that code path, set chunkDescsOffset to 0 rather than -1 in
case of "dropped more chunks from persistence than from memory" so
that no other weird things happen before the series is quarantined for
good.
The append call may reuse cds, and thus change its len.
(In practice, this wouldn't happen as cds should have len==cap.
Still, the previous order of lines was problematic.)
This decreases checkpoint size by not checkpointing things
that don't actually need checkpointing.
This is fully compatible with the v2 checkpoint format,
as it makes series appear as though the only chunksdescs
in memory are those that need persisting.
With this change the scraping caches series references and only
allocates label sets if it has to retrieve a new reference.
pkg/textparse is used to do the conditional parsing and reduce
allocations from 900B/sample to 0 in the general case.
Add metrics around checkpointing and persistence
* Add a metric to say if checkpointing is happening,
and another to track total checkpoint time and count.
This breaks the existing prometheus_local_storage_checkpoint_duration_seconds
by renaming it to prometheus_local_storage_checkpoint_last_duration_seconds
as the former name is more appropriate for a summary.
* Add metric for last checkpoint size.
* Add metric for series/chunks processed by checkpoints.
For long checkpoints it'd be useful to see how they're progressing.
* Add metric for dirty series
* Add metric for number of chunks persisted per series.
You can get the number of chunks from chunk_ops,
but not the matching number of series. This helps determine
the size of the writes being made.
* Add metric for chunks queued for persistence
Chunks created includes both chunks that'll need persistence
and chunks read in for queries. This only includes chunks created
for persistence.
* Code review comments on new persistence metrics.
When a large Prometheus starts up fresh it can take many minutes
to warmup and clear out the index queue. A larger queue means less
blocking, bigger batches and cuts down startup time by ~50%.
Keeping these around has two problems:
1) Each desc takes 64 bytes, 10 of them is 640B. This is a lot of
overhead on a 1024 byte chunk.
2) It can take well over a week to reach a point where this and thus
Prometheus memory usage as a whole enters steady state. This makes RAM
estimation very hard for users, and makes it difficult to investigate
things like memory fragmentation.
Instead we'll wipe them during each memory series maintenance cycle, and
if a query pulls them in they'll hang around as cache until the next
cycle.
Two cases:
- An unarchived metric must have at least one chunk desc loaded upon
unarchival. Otherwise, the file is gone or has size 0, which is an
inconsistency (because the series is still indexed in the archive
index). Hence, quarantining is triggered.
- If loading the chunk descs of a series with a known chunkDescsOffset
(i.e. != -1), the number of chunks loaded must be equal to
chunkDescsOffset. If not, there is a data corruption. An error is
returned, which leads to qurantining.
In any case, there is a guard added to not access the 1st element of
an empty chunkDescs slice. (That's what triggered the crashes in issue
2249.) A time series with unknown chunkDescsOffset and no chunks in
memory and no chunks on disk either could trigger that case. I would
assume such a "null series" doesn't exist, but it's not entirely
unthinkable and unreasonable to happen (perhaps in future uses of the
storage). (Create a series, and then something tries to preload chunks
before the first sample is added.)
This extracts Querier as an instantiateable and closeable object
rather than just defining extending methods of the storage interface.
This improves composability and allows abstracting query transactions,
which can be useful for transaction-level caches, consistent data views,
and encapsulating teardown.
When using the chunking code in other projects (both Weave Prism and
ChronixDB ingester), you sometimes want to know how well you are
utilizing your chunks when closing/storing them.
These more specific methods have replaced `metricForLabelMatchers`
in cases where its `map[fingerprint]metric` result type was
not necessary or was used as an intermediate step
Avoids duplicated calls to `seriesForRange` from
`QueryRange` and `QueryInstant` methods.
This is a followup to https://github.com/prometheus/prometheus/pull/2011.
This publishes more of the methods and other names of the chunk code and
moves the chunk code to its own package. There's some unavoidable
ugliness: the chunk and chunkDesc metrics are used by both packages, so
I had to move them to the chunk package. That isn't great, but I don't
see how to do it better without a larger redesign of everything. Same
for the evict requests and some other types.
* Add config, HTTP Basic Auth and TLS support to the generic write path.
- Move generic write path configuration to the config file
- Factor out config.TLSConfig -> tlf.Config translation
- Support TLSConfig for generic remote storage
- Rename Run to Start, and make it non-blocking.
- Dedupe code in httputil for TLS config.
- Make remote queue metrics global.
This is based on https://github.com/prometheus/prometheus/pull/1997.
This adds contexts to the relevant Storage methods and already passes
PromQL's new per-query context into the storage's query methods.
The immediate motivation supporting multi-tenancy in Frankenstein, but
this could also be used by Prometheus's normal local storage to support
cancellations and timeouts at some point.
CPUs have to serialise write access to a single cache line
effectively reducing level of possible parallelism. Placing
mutexes on different cache lines avoids this problem.
Most gains will be seen on NUMA servers where CPU interconnect
traffic is especially expensive
Before:
go test . -run none -bench BenchmarkFingerprintLocker
BenchmarkFingerprintLockerParallel-4 2000000 932 ns/op
BenchmarkFingerprintLockerSerial-4 30000000 49.6 ns/op
After:
go test . -run none -bench BenchmarkFingerprintLocker
BenchmarkFingerprintLockerParallel-4 3000000 569 ns/op
BenchmarkFingerprintLockerSerial-4 30000000 51.0 ns/op
My aim is to support the new grpc generic write path in Frankenstein. On the surface this seems easy - however I've hit a number of problems that make me think it might be better to not use grpc just yet.
The explanation of the problems requires a little background. At weave, traffic to frankenstein need to go through a couple of services first, for SSL and to be authenticated. So traffic goes:
internet -> frontend -> authfe -> frankenstein
- The frontend is Nginx, and adds/removes SSL. Its done this way for legacy reasons, so the certs can be managed in one place, although eventually we imagine we'll merge it with authfe. All traffic from frontend is sent to authfe.
- Authfe checks the auth tokens / cookie etc and then picks the service to forward the RPC to.
- Frankenstein accepts the reads and does the right thing with them.
First problem I hit was Nginx won't proxy http2 requests - it can accept them, but all calls downstream are http1 (see https://trac.nginx.org/nginx/ticket/923). This wasn't such a big deal, so it now looks like:
internet --(grpc/http2)--> frontend --(grpc/http1)--> authfe --(grpc/http1)--> frankenstein
Next problem was golang grpc server won't accept http1 requests (see https://groups.google.com/forum/#!topic/grpc-io/JnjCYGPMUms). It is possible to link a grpc server in with a normal go http mux, as long as the mux server is serving over SSL, as the golang http client & server won't do http2 over anything other than an SSL connection. This would require making all our service to service comms SSL. So I had a go a writing a grpc http1 server, and got pretty far. But is was a bit of a mess.
So finally I thought I'd make a separate grpc frontend for this, running in parallel with the frontend/authfe combo on a different port - and first up I'd need a grpc reverse proxy. Ideally we'd have some nice, generic reverse proxy that only knew about a map from service names -> downstream service, and didn't need to decode & re-encode every request as it went through. It seems like this can't be done with golang's grpc library - see https://github.com/mwitkow/grpc-proxy/issues/1.
And then I was surprised to find you can't do grpc from browsers! See http://www.grpc.io/faq/ - not important to us, but I'm starting to question why we decided to use grpc in the first place?
It would seem we could have most of the benefits of grpc with protos over HTTP, and this wouldn't preclude moving to grpc when its a bit more mature? In fact, the grcp FAQ even admits as much:
> Why is gRPC better than any binary blob over HTTP/2?
> This is largely what gRPC is on the wire.
This adds a flag -storage.local.engine which allows turning off local
storage in Prometheus. Instead of adding if-conditions and nil checks to
all parts of Prometheus that deal with Prometheus's local storage
(including the web interface), disabling local storage simply means
replacing the normal local storage with a noop version that throws
samples away and returns empty query results. We also don't add the noop
storage to the fanout appender to decrease internal overhead.
Instead of returning empty results, an alternate behavior could be to
return errors on any query that point out that the local storage is
disabled. Not sure which one is more preferable, so I went with the
empty result option for now.
By splitting the single queue into multiple queues and flushing each individual queue in serially (and all queues in parallel), we can guarantee to preserve the order of timestampsin samples sent to downstream systems.
- fold metric name into labels
- return initialization errors back to main
- add snappy compression
- better context handling
- pre-allocation of labels
- remove generic naming
- other cleanups
This uses a new proto format, with scope for multiple samples per
timeseries in future. This will allow users to pump samples out to
whatever they like without having to change the core Prometheus code.
There's also an example receiver to save users figuring out the
boilerplate themselves.
Turns out its valid to have an overall chunk which is smaller than the
full doubleDeltaHeaderBytes size -- if it has a single sample, it
doesn't fill the whole header. Updated unmarshalling check to respect
this.
This is (hopefully) a fix for #1653
Specifically, this makes it so that if the length for the stored
delta/doubleDelta is somehow corrupted to be too small, the attempt to
unmarshal will return an error.
The current (broken) behavior is to return a malformed chunk, which can
then lead to a panic when there is an attempt to read header values.
The referenced issue proposed creating chunks with a minimum length -- I
instead opted to just error on the attempt to unmarshal, since I'm not
clear on how it could be safe to proceed when the length is
incorrect/unknown.
The issue also talked about possibly "quarantining series", but I don't
know the surrounding code well enough to understand how to make that
happen.
Specifically, the TestSpawnNotMoreThanMaxConcurrentSendsGoroutines was failing on a fresh checkout of master.
The test had a race condition -- it would only pass if one of the
spawned goroutines happened to very quickly pull a set of samples off an
internal queue.
This patch rewrites the test so that it deterministically waits until
all samples have been pulled off that queue. In case of errors, it also
now reports on the difference between what it expected and what it found.
I verified that, if the code under test is deliberately broken, the test
successfully reports on that.
See discussion in
https://groups.google.com/forum/#!topic/prometheus-developers/bkuGbVlvQ9g
The main idea is that the user of a storage shouldn't have to deal with
fingerprints anymore, and should not need to do an individual preload
call for each metric. The storage interface needs to be made more
high-level to not expose these details.
This also makes it easier to reuse the same storage interface for remote
storages later, as fewer roundtrips are required and the fingerprint
concept doesn't work well across the network.
NOTE: this deliberately gets rid of a small optimization in the old
query Analyzer, where we dedupe instants and ranges for the same series.
This should have a minor impact, as most queries do not have multiple
selectors loading the same series (and at the same offset).
tl;dr: This is not a fundamental solution to the indexing problem
(like tindex is) but it at least avoids utilizing the intersection
problem to the greatest possible amount.
In more detail:
Imagine the following query:
nicely:aggregating:rule{job="foo",env="prod"}
While it uses a nicely aggregating recording rule (which might have a
very low cardinality), Prometheus still intersects the low number of
fingerprints for `{__name__="nicely:aggregating:rule"}` with the many
thousands of fingerprints matching `{job="foo"}` and with the millions
of fingerprints matching `{env="prod"}`. This totally innocuous query
is dead slow if the Prometheus server has a lot of time series with
the `{env="prod"}` label. Ironically, if you make the query more
complicated, it becomes blazingly fast:
nicely:aggregating:rule{job=~"foo",env=~"prod"}
Why so? Because Prometheus only intersects with non-Equal matchers if
there are no Equal matchers. That's good in this case because it
retrieves the few fingerprints for
`{__name__="nicely:aggregating:rule"}` and then starts right ahead to
retrieve the metric for those FPs and checking individually if they
match the other matchers.
This change is generalizing the idea of when to stop intersecting FPs
and go into "retrieve metrics and check them individually against
remaining matchers" mode:
- First, sort all matchers by "expected cardinality". Matchers
matching the empty string are always worst (and never used for
intersections). Equal matchers are in general consider best, but by
using some crude heuristics, we declare some better than others
(instance labels or anything that looks like a recording rule).
- Then go through the matchers until we hit a threshold of remaining
FPs in the intersection. This threshold is higher if we are already
in the non-Equal matcher area as intersection is even more expensive
here.
- Once the threshold has been reached (or we have run out of matchers
that do not match the empty string), start with "retrieve metrics
and check them individually against remaining matchers".
A beefy server at SoundCloud was spending 67% of its CPU time in index
lookups (fingerprintsForLabelPairs), serving mostly a dashboard that
is exclusively built with recording rules. With this change, it spends
only 35% in fingerprintsForLabelPairs. The CPU usage dropped from 26
cores to 18 cores. The median latency for query_range dropped from 14s
to 50ms(!). As expected, higher percentile latency didn't improve that
much because the new approach is _occasionally_ running into the worst
case while the old one was _systematically_ doing so. The 99th
percentile latency is now about as high as the median before (14s)
while it was almost twice as high before (26s).
If the chunks of a series in the checkpoint are all older then the
latest chunk on disk, the head chunk is persisted and therefore has to
be declared closed.
It would be great to have a test for this, but that would require more
plumbing, subject of #447.
PromQL only requires a much narrower interface than local.Storage in
order to run queries. Narrower interfaces are easier to replace and
test, too.
We could also change the web interface to use local.Querier, except that
we'll probably use appending functions from there in the future.
On Windows, it is not possible to rename or delete a file that is
currerntly open. This change closes the file in dropAndPersistChunks
before it tries to delete it, or rename the temporary file to it.
With a lot of series accessed in a short timeframe (by a query, a
large scrape, checkpointing, ...), there is actually quite a
significant amount of lock contention if something similar is running
at the same time.
In those cases, the number of locks needs to be increased.
On the same front, as our fingerprints don't have a lot of entropy, I
introduced some additional shuffling. With the current state, anly
changes in the least singificant bits of a FP would matter.
But only on DEBUG level.
Also, count and report the two cases of out-of-order timestamps on the
one hand and same timestamp but different value on the other hand
separately.
Before, we checkpointed after every newly detected fingerprint
collision, which is not a problem as long as collisions are
rare. However, with a sufficient number of metrics or particular
nature of the data set, there might be a lot of collisions, all to be
detected upon the first set of scrapes, and then the checkpointing
after each detection will take a quite long time (it's O(n²),
essentially).
Since we are rebuilding the fingerprint mapping during crash recovery,
the previous, very conservative approach didn't even buy us
anything. We only ever read from the checkpoint file after a clean
shutdown, so the only time we need to write the checkpoint file is
during a clean shutdown.
Prometheus is Apache 2 licensed, and most source files have the
appropriate copyright license header, but some were missing it without
apparent reason. Correct that by adding it.
The chunk encoding was hardcoded there because it mostly doesn't
matter what encoding is chosen in that test. Since type 1 is
battle-hardened enough, I'm switching to type 2 here so that we can
catch unexpected problems as a byproduct. My expectation is that the
chunk encoding doesn't matter anyway, as said, but then "unexpected
problems" contains the word "unexpected".
So far, the last sample in a chunk was saved twice. That's required
for adding more samples as we need to know the last sample added to
add more samples without iterating through the whole chunk. However,
once the last sample was added to the chunk before it's full, there is
no need to save it twice. Thus, the very last sample added to a chunk
can _only_ be saved in the header fields for the last sample. The
chunk has to be identifiable as closed, then. This information has
been added to the flags byte.
This improves fuzz testing in two ways:
(1) More realistic time stamps. So far, the most common case in
practice was very rare in the test: Completely regular increases of
the timestamp.
(2) Verify samples by scanning through the whole relevant section of
the series.
For Gorilla-like chunks, this showed two things:
(1) With more regularly increasing time stamps, BenchmarkFuzz is
essentially as fast as with the traditional chunks:
```
BenchmarkFuzzChunkType0-8 2 972514684 ns/op 83426196 B/op 2500044 allocs/op
BenchmarkFuzzChunkType1-8 2 971478001 ns/op 82874660 B/op 2512364 allocs/op
BenchmarkFuzzChunkType2-8 2 999339453 ns/op 76670636 B/op 2366116 allocs/op
```
(2) There was a bug related to when and how the chunk footer is
overwritten to make use for the last sample. This wasn't exposed by
random access as the last sample of a chunk is retrieved from the
values in the header in that case.
This is not a verbatim implementation of the Gorilla encoding. First
of all, it could not, even if we wanted, because Prometheus has a
different chunking model (constant size, not constant time). Second,
this adds a number of changes that improve the encoding in general or
at least for the specific use case of Prometheus (and are partially
only possible in the context of Prometheus). See comments in the code
for details.
It is now also used in label matching, so the name of the metric
changed from `prometheus_local_storage_invalid_preload_requests_total`
to `non_existent_series_matches_total'.
Only return an error where callers are doing something with it except
simply logging and ignoring.
All the errors touched in this commit flag the storage as dirty
anyway, and that fact is logged anyway. So most of what is being
removed here is just log spam.
As discussed earlier, the class of errors that flags the storage as
dirty signals fundamental corruption, no even bubbling up a one-time
warning to the user (e.g. about incomplete results) isn't helping much
because _anything_ happening in the storage has to be doubted from
that point on (and in fact retroactively into the past, too). Flagging
the storage dirty, and alerting on it (plus marking the state in the
web UI) is the only way I can see right now.
As a byproduct, I cleaned up the setDirty method a bit and improved
the logged errors.
WIP: This needs more tests.
It now gets a from and through value, which it may opportunistically
use to optimize the retrieval. With possible future range indices,
this could be used in a very efficient way. This change merely applies
some easy checks, which should nevertheless solve the use case of
heavy rule evaluations on servers with a lot of series churn.
Idea is the following:
- Only archive series that are at least as old as the headChunkTimeout
(which was already extremely unlikely to happen).
- Then maintain a high watermark for the last archival, i.e. no
archived series has a sample more recent than that watermark.
- Any query that doesn't reach to a time before that watermark doesn't
have to touch the archive index at all. (A production server at
Soundcloud with the aforementioned series churn and heavy rule
evaluations spends 50% of its CPU time in archive index
lookups. Since rule evaluations usually only touch very recent
values, most of those lookup should disappear with this change.)
- Federation with a very broad label matcher will profit from this,
too.
As a byproduct, the un-needed MetricForFingerprint method was removed
from the Storage interface.
This finally extracts all the common code of the two chunk iterators
into one. Any future chunk encodings with fast access by index can use
the same iterator by simply providing an indexAccessor. Other future
chunk encodings without fast index access (like Gorilla-style) can
still implement the chunkIterator interface as usual.
For one, remove unneeded methods.
Then, instead of using a channel for all values, use a
bufio.Scanner-like interface. This removes the need for creating a
goroutine and avoids the (unnecessary) locking performed by channel
sending and receiving.
This will make it much easier to write new chunk implementations (like
Gorilla-style encoding).
I needed this today for debugging. It can certainly be improved, but
it's already quite helpful.
I refactored the reading of heads.db files out of persistence, which
is an improvement, too.
I made minor changes to the cli package to allow outputting via the
io.Writer interface.
Obviously, it's really bad to depend on timing here. The proper fix
would be to have something like WaitForIndexing for other things to
wait for, too.
For now, let's see if the wait time increase fixes the issue.
This fixes https://github.com/prometheus/prometheus/issues/1059 , but
not in the obvious way (simply not updating the persist watermark,
because that's actually not that simple - we don't really know what
has gone wrong exactly). As any errors relevant here are most likely
caused by severe and unrecoverable problems with the series file,
Using the now quarantine feature is the right step. We don't really
have to be worried about any inconsistent state of the series because
it will be removed for good ASAP. Another plus is that we don't have
to declare the whole storage dirty anymore.
This requires all the panic calls upon unexpected data to be converted
into errors returned. This pollute the function signatures quite
lot. Well, this is Go...
The ideas behind this are the following:
- panic only if it's a programming error. Data corruptions happen, and
they are not programming errors.
- If we detect a data corruption, we "quarantine" the series,
essentially removing it from the database and putting its data into
a separate directory for forensics.
- Failure during writing to a series file is not considered corruption
automatically. It will call setDirty, though, so that a
crashrecovery upon the next restart will commence and check for
that.
- Series quarantining and setDirty calls are logged and counted in
metrics, but are hidden from the user of the interfaces in
interface.go, whith the notable exception of Append(). The reasoning
is that we treat corruption by removing the corrupted series, i.e. a
query for it will return no results on its next call anyway, so
return no results right now. In the case of Append(), we want to
tell the user that no data has been appended, though.
Minor side effects:
- Now consistently using filepath.* instead of path.*.
- Introduced structured logging where I touched it. This makes things
less consistent, but a complete change to structured logging would
be out of scope for this PR.
Fixes https://github.com/prometheus/prometheus/issues/1401
This remove the last (and in fact bogus) use of BoundaryValues.
Thus, a whole lot of unused (and arguably sub-optimal / ugly) code can
be removed here, too.
In a way, our instants were also ranges, just with the staleness delta
as range length. They are no treated equally, just that in one case,
the range length is set as range, in the other the staleness
delta. However, there are "real" instants where start and and time of
a query is the same. In those cases, we only want to return a single
value (the one closest before or at the equal start and end time). If
that value is the last sample in the series, odds are we have it
already in the series object. In that case, there is no need to pin or
load any chunks. A special singleSampleSeriesIterator is created for
that. This should greatly speed up instant queries as they happen
frequently for rule evaluations.
This implies a slight change of behavior as only samples added to the
respective instance of a memorySeries are returned. However, this is
most likely anyway what we want.
Following cases:
- Server has been restarted: Given the time it takes to cleanly
shutdown and start up a server, the series are now stale anyway. An
improved staleness handling (still to be implemented) will be based
on tracking if a given target is continuing to expose samples for a
given time series. In that case, we need a full scrape cycle to
decide about staleness. So again, it makes sense to consider
everything stale directly after a server restart.
- Series unarchived due to a read request: The series is definitely
stale so we don't want to return anything anyway.
- Freshly created time series or series unarchived because of a sample
append: That happens because appending a sample is imminent. Before
the fingerprint lock is released, the series will have received a
sample, and lastSamplePair will always returned the expected value.
Formalize ZeroSamplePair as return value for non-existing samples.
Change LastSamplePairForFingerprint to return a SamplePair (and not a
pointer to it), which saves allocations in a potentially extremely
frequent call.
This will fix issue #1035 and will also help to make issue #1264 less
bad.
The fundamental problem in the current code:
In the preload phase, we quite accurately determine which chunks will
be used for the query being executed. However, in the subsequent step
of creating series iterators, the created iterators are referencing
_all_ in-memory chunks in their series, even the un-pinned ones. In
iterator creation, we copy a pointer to each in-memory chunk of a
series into the iterator. While this creates a certain amount of
allocation churn, the worst thing about it is that copying the chunk
pointer out of the chunkDesc requires a mutex acquisition. (Remember
that the iterator will also reference un-pinned chunks, so we need to
acquire the mutex to protect against concurrent eviction.) The worst
case happens if a series doesn't even contain any relevant samples for
the query time range. We notice that during preloading but then we
will still create a series iterator for it. But even for series that
do contain relevant samples, the overhead is quite bad for instant
queries that retrieve a single sample from each series, but still go
through all the effort of series iterator creation. All of that is
particularly bad if a series has many in-memory chunks.
This commit addresses the problem from two sides:
First, it merges preloading and iterator creation into one step,
i.e. the preload call returns an iterator for exactly the preloaded
chunks.
Second, the required mutex acquisition in chunkDesc has been greatly
reduced. That was enabled by a side effect of the first step, which is
that the iterator is only referencing pinned chunks, so there is no
risk of concurrent eviction anymore, and chunks can be accessed
without mutex acquisition.
To simplify the code changes for the above, the long-planned change of
ValueAtTime to ValueAtOrBefore time was performed at the same
time. (It should have been done first, but it kind of accidentally
happened while I was in the middle of writing the series iterator
changes. Sorry for that.) So far, we actively filtered the up to two
values that were returned by ValueAtTime, i.e. we invested work to
retrieve up to two values, and then we invested more work to throw one
of them away.
The SeriesIterator.BoundaryValues method can be removed once #1401 is
fixed. But I really didn't want to load even more changes into this
PR.
Benchmarks:
The BenchmarkFuzz.* benchmarks run 83% faster (i.e. about six times
faster) and allocate 95% fewer bytes. The reason for that is that the
benchmark reads one sample after another from the time series and
creates a new series iterator for each sample read.
To find out how much these improvements matter in practice, I have
mirrored a beefy Prometheus server at SoundCloud that suffers from
both issues #1035 and #1264. To reach steady state that would be
comparable, the server needs to run for 15d. So far, it has run for
1d. The test server currently has only half as many memory time series
and 60% of the memory chunks the main server has. The 90th percentile
rule evaluation cycle time is ~11s on the main server and only ~3s on
the test server. However, these numbers might get much closer over
time.
In addition to performance improvements, this commit removes about 150
LOC.
The First time is kind of trivial as we always know it when we create
a new chunkDesc.
The last time is only know when the chunk is closed, so we have to set
it at that time.
The change saves a lot of digging down into the chunk
itself. Especially the last time is relative expensive as it involves
the creation of an iterator. The first time access now doesn't require
locking, which is also a nice gain.
Chris Marchbanks