Fixes most of https://github.com/prometheus/prometheus/issues/15202
This should address all areas of the UI except for the autocompletion in the
codemirror-promql text editor. The strategy here is that any time we print or
internally serialize (like for the PromLens tree view) either a metric name or
a label name as part of a selector or in other relevant parts of PromQL, we
check whether it contains characters beyond what was previously supported, and
if so, quote and escape it. In the case of metric names, we also have to move
them from the beginning of the selector into the curly braces.
Signed-off-by: Julius Volz <julius.volz@gmail.com>
The `info` function is an experiment to improve UX
around including labels from info metrics.
`info` has to be enabled via the feature flag `--enable-feature=promql-experimental-functions`.
This MVP of info simplifies the implementation by assuming:
* Only support for the target_info metric
* That target_info's identifying labels are job and instance
Also:
* Encode info samples' original timestamp as sample value
* Deduce info series select hints from top-most VectorSelector
---------
Signed-off-by: Arve Knudsen <arve.knudsen@gmail.com>
Co-authored-by: Ying WANG <ying.wang@grafana.com>
Co-authored-by: Augustin Husson <augustin.husson@amadeus.com>
Co-authored-by: Bartlomiej Plotka <bwplotka@gmail.com>
Co-authored-by: Björn Rabenstein <github@rabenste.in>
Co-authored-by: Bryan Boreham <bjboreham@gmail.com>
For: #14355
This commit updates Prometheus to adopt stdlib's log/slog package in
favor of go-kit/log. As part of converting to use slog, several other
related changes are required to get prometheus working, including:
- removed unused logging util func `RateLimit()`
- forward ported the util/logging/Deduper logging by implementing a small custom slog.Handler that does the deduping before chaining log calls to the underlying real slog.Logger
- move some of the json file logging functionality to use prom/common package functionality
- refactored some of the new json file logging for scraping
- changes to promql.QueryLogger interface to swap out logging methods for relevant slog sugar wrappers
- updated lots of tests that used/replicated custom logging functionality, attempting to keep the logical goal of the tests consistent after the transition
- added a healthy amount of `if logger == nil { $makeLogger }` type conditional checks amongst various functions where none were provided -- old code that used the go-kit/log.Logger interface had several places where there were nil references when trying to use functions like `With()` to add keyvals on the new *slog.Logger type
Signed-off-by: TJ Hoplock <t.hoplock@gmail.com>
Some of the changes are a bit unreadable because the previous files were not
saved with the project's linter / auto-formatter settings applied. But it's
basically:
* For icons that are not Mantine-native components, use the rem() function
for computing their size, so they scale correctly with the root font size.
See https://mantine.dev/styles/rem/.
* Try a different icon for the notifications tray, since the bell icon was
already used for Prometheus alerts. Other candidates from
https://tabler.io/icons would be IconExclamationCircle or
IconDeviceDesktopExclamation or IconMessageCircleExclamation.
* The server startup alert looked a bit cramped, introduced a Stack to add
spacing between the text and the progress bar.
* Added a bit of spacing between notification text and date. Things looked
cramped. To make things look ok with that, I also top-aligned the
notification text and icon.
Signed-off-by: Julius Volz <julius.volz@gmail.com>
api: Improve doc comments for v1.MinTime and v1.MaxTime
While investigated something mostly unrelated, I got nerd-sniped by
the calculation of v1.MinTime and v1.MaxTime. The seemingly magic
number in there (62135596801) needed an explanation. While looking for
it, I found out that the offsets used here are actually needlessly
conservative. Since the timestamps are so far in the past or future,
respectively, that there is no practical impact, except that the
calculation is needlessly obfuscated. However, we won't change the
values now to not cause any confusion for users of this code. Still, I
think the doc comment should explain the circumstances so nobody gets
nerd-sniped again as I did today.
For the record: 62135596800 is the difference in seconds between 0001-01-01 00:00:00 (Unix time zero point) and 1971-01-01 00:00:00 (Go time zero point) in the Gregorian calendar. If "Prometheus time" were in seconds (not milliseconds), that difference would be relevant to prevent over-/underflow when converting from "Prometheus time" to "Go time".
Signed-off-by: beorn7 <beorn@grafana.com>
---------
Signed-off-by: beorn7 <beorn@grafana.com>
The OTLP receiver can now considered stable. We've had it for longer
than a year in main and has received constant improvements.
Signed-off-by: Jesus Vazquez <jesusvzpg@gmail.com>
Without this, the page that is shown first renders once with an empty path
prefix value, since the settings update takes a render cycle to complete.
However, we only fetch certain data from the API exactly once for a given
page, and not for every re-render with changed path prefix value (and we
also wouldn't want to fetch it from the wrong location initially).
This duplicates the served endpoint list once more, but exporting them from
App.tsx would also have been dirty (hot reload only works when a file only
exports one component and nothing else, thus there'd be a linter warning).
Signed-off-by: Julius Volz <julius.volz@gmail.com>
This switches from the prehistoric EventSource API to the more modern
fetch-event-source package. That packages gives us full control over the
retries.
It also gives us the opportunity to close the event source when the
browser tab is hidden, saving resources.
Signed-off-by: Julien <roidelapluie@o11y.eu>
This commit introduces a new `/api/v1/notifications/live` endpoint that
utilizes Server-Sent Events (SSE) to stream notifications to the web UI.
This is used to display alerts such as when a configuration reload
has failed.
I opted for SSE over WebSockets because SSE is simpler to implement and
more robust for our use case. Since we only need one-way communication
from the server to the client, SSE fits perfectly without the overhead
of establishing and maintaining a two-way WebSocket connection.
When the SSE connection fails, we go back to a classic
/api/v1/notifications API endpoint.
This commit also contains the required UI changes for the new Mantine UI.
Signed-off-by: Julien <roidelapluie@o11y.eu>
Previously, scrapes durations that are very short (e.g., connection refused)
could show as empty (durations under 1 millisecond).
This commit ensures that sub-millisecond durations are correctly
displayed as "0ms" or "1ms" when necessary.
- Adjusted `humanizeDuration` to round sub-millisecond durations to the
nearest millisecond.
- Updated unit tests to verify the correct handling of sub-millisecond
values.
Signed-off-by: Julien <roidelapluie@o11y.eu>
Jan Fajerski