GitVita
New repositoryImport repositoryNew issue
Your repositories Your dashboard Sign in

cloudflare/pint

Public

mirrored from https://github.com/cloudflare/pintAvailable

Watch0Fork0Star0
CodeCommitsIssuesPull requestsActionsInsightsSecurity
v0.75.0

Branches

Tags

  • No tags available.
0Branches0Tags
Go to file
Add file
Code

Clone

HTTPS

Download ZIP

pint/docs

docs/changelog.md

2234lines · modecode

RawDownload
Latest commit unavailable.
1# Changelog
2
3## v0.75.0
4
5### Added
6
7- Added [promql/selector](checks/promql/selector.md) check - [#1416](https://github.com/cloudflare/pint/issues/1416)
8 and [#1540](https://github.com/cloudflare/pint/issues/1540).
9
10## v0.74.8
11
12### Fixed
13
14- Fixed incorrect suggestions from [query/cost](checks/query/cost.md) for function calls with different arguments.
15
16## v0.74.7
17
18### Fixed
19
20- Fixed a bug in GitLab integration causing problems fail to create comments when they are reported on deleted lines.
21- Fixed incorrect suggestions from [query/cost](checks/query/cost.md) for rules joining on multiple metrics.
22
23## v0.74.6
24
25### Fixed
26
27- Fixed a bug in GitLab integration causing some problems fail to create comments when files were renamed and modified.
28
29## v0.74.5
30
31### Fixed
32
33- Fixed incorrect incorrect suggestions generated by [query/cost](checks/query/cost.md) check.
34- Fixed unmodified lines incorrectly marked as modified when running `pint ci`.
35
36## v0.74.4
37
38### Fixed
39
40- Fixed false positive reports from [promql/rate](checks/promql/rate.md) check when native histograms or summaries are used.
41
42## v0.74.3
43
44### Fixed
45
46- Fixed false positive reports from [promql/fragile](checks/promql/fragile.md) check - [#1466](https://github.com/cloudflare/pint/issues/1466).
47- Fixed incorrect reports from [query/cost](checks/query/cost.md) when multiple rules with `rate(...)` were present
48 and each had a different rate time window.
49
50## v0.74.2
51
52### Fixed
53
54- Fixed an issue where pint would fail to create MR comment if GitLab API returned empty diff body.
55
56## v0.74.1
57
58### Fixed
59
60- Fixed incorrect suggestion from [query/cost](checks/query/cost.md) when a query uses joins or unless.
61
62## v0.74.0
63
64### Added
65
66- [promql/fragile](checks/promql/fragile.md) check will now warn about alerting rules using
67 aggregations that might cause false positives when Prometheus restarts.
68- [query/cost](checks/query/cost.md) check will now try to find queries that perform operations
69 for which there is a recording rule, and suggest using these recording rules instead to speed
70 up the query.
71
72### Fixed
73
74- Improved accuracy of messages produced by the [promql/impossible](checks/promql/impossible.md) check.
75
76## v0.73.7
77
78### Fixed
79
80- Fixed a bug in GitLab integration causing incorrect line numbers to be used for comments.
81
82## v0.73.6
83
84### Fixed
85
86- Fixed a bug in GitLab integration causing some problems fail to create comments when files were renamed.
87
88## v0.73.5
89
90### Fixed
91
92- Fixed incorrect error message for some problems reported by [alerts/template](checks/alerts/template.md) check.
93
94## v0.73.4
95
96### Fixed
97
98- Fixed handling of resolved comments when using GitLab with `pint ci` command.
99
100## v0.73.3
101
102### Fixed
103
104- Fixed incorrect value in `This pint run would create N comment(s), which is more than the limit ...` comments
105 when using GitHub or GitLab with `pint ci` command.
106
107## v0.73.2
108
109### Fixed
110
111- Fixed handling of stale comments when using GitLab with `pint ci` command.
112
113## v0.73.1
114
115### Fixed
116
117- Fixed `relaxed` parsing mode not finding all rules in some files - [#1405](https://github.com/cloudflare/pint/issues/1405).
118
119## v0.73.0
120
121### Changed
122
123- When pint finds the exact same problem generating multiple problem reports it will now hide duplicated reports.
124 There is a new flag `pint --show-duplicates ...` that you can use to show all instances of each problem.
125 For example if we have three different rules with the same smelly selector:
126
127 ```yaml
128 groups:
129 - name: foo
130 rules:
131 - record: events:a
132 expr: events_total{instance="a", job=~"foo.+"}
133 - record: events:b
134 expr: events_total{instance="b", job=~"foo.+"}
135 - record: events:c
136 expr: events_total{instance="c", job=~"foo.+"}
137 ```
138
139 Previous versions of pint would print each problem:
140
141 ```yaml
142 Warning: smelly regexp selector (promql/regexp)
143 ---> rules/1.yaml:5 -> `events:a`
144 5 | expr: events_total{instance="a", job=~"foo.+"}
145 ^^^^^^^^^^^^ `{job=~"foo.+"}` looks like a smelly selector ...
146
147 Warning: smelly regexp selector (promql/regexp)
148 ---> rules/1.yaml:7 -> `events:b`
149 7 | expr: events_total{instance="b", job=~"foo.+"}
150 ^^^^^^^^^^^^ `{job=~"foo.+"}` looks like a smelly selector ...
151
152 Warning: smelly regexp selector (promql/regexp)
153 ---> rules/1.yaml:9 -> `events:c`
154 9 | expr: events_total{instance="c", job=~"foo.+"}
155 ^^^^^^^^^^^^ `{job=~"foo.+"}` looks like a smelly selector ...
156 ```
157
158 This release of pint will (by default) show it only once:
159
160 ```yaml
161 Warning: smelly regexp selector (promql/regexp)
162 ---> rules/1.yaml:5 -> `events:a` [+2 duplicates]
163 5 | expr: events_total{instance="a", job=~"foo.+"}
164 ^^^^^^^^^^^^ `{job=~"foo.+"}` looks like a smelly selector ...
165 ```
166
167## v0.72.1
168
169### Fixed
170
171- `pint_problem` metrics exposed when running `pint watch` were lacking problem details.
172
173## v0.72.0
174
175### Added
176
177- Labels set on [the group level](https://prometheus.io/docs/prometheus/latest/configuration/recording_rules/#rule_group) are now supported by pint.
178- [promql/series](checks/promql/series.md) check now supports `snooze` comments with a metric selector.
179
180### Fixed
181
182- [alerts/comparison](checks/alerts/comparison.md) was refactored to output more accurate reports.
183- [alerts/template](checks/alerts/template.md) was refactored to output more accurate reports.
184- [promql/vector_matching](checks/promql/vector_matching.md) will now produce more accurate messages with more details.
185
186## v0.71.8
187
188### Fixed
189
190- Don't try to report problems to BitBucket for unmodified rules.
191
192## v0.71.7
193
194### Fixed
195
196- Fixed a panic when parsing empty documents - [#1355](https://github.com/cloudflare/pint/issues/1355).
197
198## v0.71.6
199
200### Fixed
201
202- [promql/series](checks/promql/series.md) check will now output more accurate comments.
203
204## v0.71.5
205
206### Fixed
207
208- [alerts/template](checks/alerts/template.md) check will now output more accurate comments.
209
210## v0.71.4
211
212### Fixed
213
214- Fix [promql/series](checks/promql/series.md) to not fail on `absent({...})` calls.
215- Fix formatting of BitBucket comments.
216
217## v0.71.3
218
219### Fixed
220
221- Some parts of comments reported to GitHub & GitLab were not using collapsible blocks as intended.
222
223## v0.71.2
224
225### Fixed
226
227- Print only relevant lines when reporting problems.
228- Make comments reported to GitHub & GitLab more compact by using collapsible blocks.
229
230## v0.71.1
231
232### Fixed
233
234- Don't create empty markdown blocks when reporting problems to GitHub & GitLab.
235- Added missing problem details to comments reported to BitBucket.
236
237## v0.71.0
238
239### Added
240
241- Added `names` option to the `parser` block, which controls how does Prometheus validates
242 label names.
243- Added [promql/impossible](checks/promql/impossible.md) check. This check duplicates some
244 of the functionality of the [promql/vector_matching](checks/promql/vector_matching.md) so
245 in the future the scope of [promql/vector_matching](checks/promql/vector_matching.md) checks
246 will be reduced to avoid duplicated reports for the same issue.
247- Added `ignoreMatchingElsewhere` option to [promql/series](checks/promql/series.md) check.
248
249### Changed
250
251- Removed aggregation checks from the [promql/fragile](checks/promql/fragile.md) check - #1289.
252- [promql/regexp](checks/promql/regexp.md) check severity was reduced from `bug` to a `warning`.
253- This release contains some major changes to the way problems are reported by pint.
254 Each problem might now point to the specific part of the PromQL query that caused the issue
255 rather than just the line range on which the query is defined. Most, but not all, checks
256 were updated to emit more detailed problem descriptions.
257
258### Fixed
259
260- Improved the logic of [promql/series](checks/promql/series.md) check to skip checks on some selectors
261 which absence might be expected. For example using `foo unless bar` will now only have `foo`
262 tested while `bar` is ignored by this check. This is because the query implies that `bar`
263 might be present or missing and depending on that `foo` is evaluated.
264- Don't report regexp matchers with only digits in them (`{code=~"5.*"}`) as smelly in [promql/regexp](checks/promql/regexp.md) check.
265
266## v0.70.0
267
268### Added
269
270- When running `pint watch` command `/health` HTTP endpoint can now be used for liveness probes.
271
272### Changed
273
274- When pint runs online checks against Thanos or other service that implements some, but not all, Prometheus APIs
275 it would reports errors due to some API requests returning `404 Not Found` HTTP errors.
276 Now pint will automatically disable checks that rely on Prometheus API endpoints that are not supported
277 and create a summary comment (when running `pint ci`) that lists disabled checks.
278 This applies to these API endpoints:
279
280 - `/api/v1/status/config`
281 - `/api/v1/status/flags`
282 - `/api/v1/metadata`
283
284## v0.69.1
285
286### Fixed
287
288- `# pint file/owner` comments were not validated properly for files with no rules.
289 This is now fixed.
290
291## v0.69.0
292
293### Added
294
295- Added `schema` option to the `parser` configuration block for setting rule validation
296 schema. This option is only used when files are parsed in strict mode - which is when
297 rule file path does NOT match any of the `parser:relaxed` regex values or when simply
298 `parser:relaxed` is not set at all.
299 Default value is `prometheus` and tells pint to expect rule files with the schema
300 expected by Prometheus itself. If you use pint to validate rules loaded into Thanos Rule
301 component then set `schema` to `thanos` in your pint config file:
302
303 ```js
304 parser {
305 schema = "thanos"
306 }
307 ```
308
309 File schema when using `schema: prometheus` (default):
310
311 ```yaml
312 groups:
313 - name: example
314 rules:
315 - record: ...
316 expr: ...
317 - alert: ...
318 expr: ...
319 ```
320
321 When using `schema: thanos`:
322
323 ```yaml
324 groups:
325 - name: example
326 partial_response_strategy: abort
327 rules:
328 - record: ...
329 expr: ...
330 - alert: ...
331 expr: ...
332 ```
333
334- Rules configured in `pint` config can now be locked - when a rule is locked it cannot
335 be disabled by users by adding a `# pint disable ...` or `# pint snooze ...` comments.
336
337### Fixed
338
339- The console reporter won't color the output if `--no-color` flag is set.
340
341## v0.68.0
342
343### Added
344
345- Added [rule/report](checks/rule/report.md) check.
346
347### Changed
348
349- pint now uses [Prometheus 3.0](https://prometheus.io/blog/2024/11/14/prometheus-3-0/) libraries
350 for parsing PromQL, which adds support for new query syntax that allows for dots and UTF-8 chars
351 in metric/label names, example:
352
353 ```js
354 {"status.üp"} == 0
355 ```
356
357- [promql/rate](checks/promql/rate.md) will now report a warning if it detects a `rate(sum(...))`
358 but doesn't have metadata to confirm if `...` is a counter or not.
359
360## v0.67.3
361
362### Fixed
363
364- Fixed a crash when parsing `vector()` calls with non-number arguments.
365
366## v0.67.2
367
368### Fixed
369
370- Improved the accuracy of [alerts/template](checks/alerts/template.md) check.
371
372## v0.67.1
373
374### Fixed
375
376- Improved message formatting in the [alerts/template](checks/alerts/template.md) check.
377- Fixed problem reports for multi line queries.
378
379## v0.67.0
380
381### Added
382
383- Added `--checkstyle` flag to `pint lint` & `pint ci` for writing XML report file
384 in `checkstyle` format - [#1129](https://github.com/cloudflare/pint/pull/1129).
385- Added `--json` flag to both `pint lint` and `pint ci` commands, this enables writing a JSON file
386 with the report of all problems [#606](https://github.com/cloudflare/pint/issues/606).
387- [promql/fragile](checks/promql/fragile.md) will now warn when alerting rules are using
388 one of the aggregation operation that can return different series on every evaluation,
389 which can cause alert floppiness - [#820](https://github.com/cloudflare/pint/issues/820).
390- [promql/regexp](checks/promql/regexp.md) check now supports extra configuration options
391 to disable reports on smelly selector - [#1096](https://github.com/cloudflare/pint/issues/1096).
392- Checks can be enabled or disabled specifically for some Prometheus rules via `rule {}` config blocks.
393 Adding `enable` or `disable` option with a list of checks names allows to selectively enable or disable
394 checks only for Prometheus rules that match given `rule {}` definition.
395 Enabling checks only for matching rules will only work if these checks are disabled globally via
396 `check { disabled = [] }` config block.
397 For example to disable `promql/rate` check for all rules except alerting rules in the `rules/critical` folder:
398
399 ```js
400 checks {
401 // This will disable promql/rate by default.
402 disabled = [ "promql/rate" ]
403 }
404 rule {
405 match {
406 path = "rules/critical/.*"
407 kind = "alerting"
408 }
409 // This will enable promql/rate only for Prometheus rules matching all our match conditions above.
410 enable = [ "promql/rate" ]
411 }
412 ```
413
414### Changed
415
416- [alerts/template](checks/alerts/template.md) check was refactored and will now produce more accurate results.
417 Messages produced by this check might include details of the PromQL query fragment causing the problem
418 if the query is complex enough.
419
420### Fixed
421
422- Don't try to create GitLab comments on unmodified lines - [#1147](https://github.com/cloudflare/pint/pull/1147).
423
424## v0.66.1
425
426### Fixed
427
428- Fixed message formatting in the [promql/series](checks/promql/series.md) check.
429
430## v0.66.0
431
432### Added
433
434- Added `fallbackTimeout` option to the [promql/series](checks/promql/series.md) check
435 that controls how much time pint can spend checking other Prometheus servers for missing
436 metrics.
437
438### Fixed
439
440- Reverted `Fixed colored output on some environments - #1106` change
441 as it was breaking GitHub report comments.
442- Fixed panics in [rule/duplicate](checks/rule/duplicate.md) check.
443
444## v0.65.3
445
446### Fixed
447
448- Fixed GitHub actions permissions.
449
450## v0.65.2
451
452### Fixed
453
454- Improved accuracy of the [rule/duplicate](checks/rule/duplicate.md) check.
455- Fixed GitHub reporter trying to create pull request comments to unmodified lines - #1120.
456- Fixed colored output on some environments - #1106.
457- Show correct line number when reporting YAML syntax errors.
458
459## v0.65.1
460
461### Fixed
462
463- Fixed a bug in `match` block `state` handling that caused all rules to always match any state.
464
465## v0.65.0
466
467### Added
468
469- [promql/regexp](checks/promql/regexp.md) check will now look for smelly regexp selectors.
470 See check docs for details.
471- [promql/range_query](checks/promql/range_query.md) now allows to configure a custom maximum
472 duration for range queries - #1064.
473- Added `--enabled` flag to the pint command. Passing this flag will only run selected check(s).
474- Added `state` option to the rule `match` block. See [configuration](configuration.md) docs for details.
475
476### Fixed
477
478- Don't try to report problem on unmodified files when using GitHub reporter.
479- If there is a pint config file present then pint will now always add it to the `parser` block `exclude` list.
480 This is to avoid trying to parse it as a rule file if it's included in the same folder as rules.
481
482## v0.64.1
483
484### Fixed
485
486- Validate `ignoreLabelsValue` option values in the pint config.
487
488## v0.64.0
489
490### Added
491
492- Added `ignoreLabelsValue` to [promql/series](checks/promql/series.md) check settings.
493- Added `include` & `exclude` options were moved from the `ci` to the `parser` configuration block.
494 They now apply to all pint commands, not just `pint ci` - #631.
495
496## v0.63.0
497
498### Added
499
500- Added [alerts/absent](checks/alerts/absent.md) check.
501- Added [rule/name](checks/rule/name.md) check - #1020.
502
503### Changed
504
505- [promql/vector_matching](checks/promql/vector_matching.md) will now report more details, including which Prometheus server reports problems and which part of the query is the issue.
506- GitHub report code was refactored, it should behave as before.
507- When running `pint ci` pint will now run all checks on unmodified rules when a `file/disable` comment was removed.
508
509### Fixed
510
511- Fixed false positive warnings from [alerts/comparison](checks/alerts/comparison.md) when using `absent_over_time()`.
512- GitHub reporter will now wait before making more requests if it's rate limited - #699.
513
514## v0.62.2
515
516### Fixed
517
518- When using BitBucket reporter `pint ci` might create a comment longer than the limit allowed by BitBucket.
519 To avoid this pint will now truncate long comments.
520
521## v0.62.1
522
523### Fixed
524
525- Fixed false positive warnings from [rule/dependency](checks/rule/dependency.md) check.
526
527## v0.62.0
528
529### Added
530
531- [promql/series](checks/promql/series.md) check will now generate warnings if there are `# pint disable`
532 or `# pint rule/set` comments that are not matching any valid query selector or Prometheus server.
533
534### Fixed
535
536- [promql/series](checks/promql/series.md) will now parse `rule/set` comments that target specific
537 time series selectors in PromQL the same way as `# pint disable` comments do.
538
539## v0.61.2
540
541### Fixed
542
543- Fixed false positive reports about removed rules when running `pint ci` on a branch that
544 contains YAML syntax errors.
545
546## v0.61.1
547
548### Fixed
549
550- Fixed release workflow on GitHub.
551
552## v0.61.0
553
554### Fixed
555
556- Don't suggest using `humanize` when alert template is already using printf on format the `$value`.
557- Fixed git history parsing when running `pint ci` on a branch that include merge commits.
558
559### Changed
560
561- Refactored YAML syntax checks to avoid using [rulefmt.Parse](https://pkg.go.dev/github.com/prometheus/prometheus@v0.53.0/model/rulefmt#Parse) and effectively parsing rules twice. Some error messages will have different formatting.
562
563## v0.60.0
564
565### Fixed
566
567- [alerts/annotations](checks/alerts/annotation.md) and [rule/label](checks/rule/label.md) will
568 now report when a label or annotation is required but set to an empty value.
569- Fixed handling of `# pint ignore/line` comments on line including multiple `#` characters.
570- When reporting problems all messages will now use `publicURI` from each `prometheus` definition.
571
572## v0.59.0
573
574### Added
575
576- Added GitLab support when running `pint ci`.
577
578### Changed
579
580- Reduced severity of problems reported by [promql/counter](checks/promql/counter.md) due
581 to high number of false positives. It will now report warnings instead of errors.
582
583### Fixed
584
585- Fixed false positive reports from [rule/duplicate](checks/rule/duplicate.md) when a rule file
586 is being deleted and the rules are moved to a different file.
587
588## v0.58.1
589
590### Fixed
591
592- Correctly handle default Prometheus retention value in [promql/range_query](checks/promql/range_query.md)
593 check - #958.
594
595## v0.58.0
596
597### Fixed
598
599- Fixed false positives from [rule/duplicate](checks/rule/duplicate.md) check
600 when running `pint ci` on files that are both edited and renamed in the same PR.
601- [rule/dependency](checks/rule/dependency.md) will no longer report removed symlinks.
602
603## v0.57.3
604
605### Fixed
606
607- Fixed compatibility with older git releases.
608- Fixed `maxComments` defaulting to zero when using GitHub reporter.
609- Fixed `maxComments` comment creation on GitHub - #935.
610
611## v0.57.2
612
613### Fixed
614
615- Fixed release workflow on GitHub actions.
616
617## v0.57.0
618
619### Added
620
621- GitHub and BitBucket reporters now supports `maxComments` option to limit the number
622 of comments pint can create on a single pull request.
623 Default value of `maxComments` is `50`.
624
625## v0.56.2
626
627### Fixed
628
629- [query/cost](checks/query/cost.md) will now only create reports if a query is more
630 expensive than any of the configured limits.
631- Fixed duplicated reports when using BitBucket reporter.
632
633## v0.56.1
634
635### Fixed
636
637- [promql/counter](checks/promql/counter.md) check will now consider a metric
638 to be a counter only if all metadata entries for it use `TYPE counter`.
639 Previously it would check for at least one metadata entry with `TYPE counter`.
640
641## v0.56.0
642
643### Added
644
645- Added [promql/counter](checks/promql/counter.md) check - #655.
646- Check using custom rules now accept an optional `comment` option for adding
647 a text comment to all reported problems.
648 Checks supporting `comment` option:
649 - [alerts/annotation](checks/alerts/annotation.md)
650 - [alerts/count](checks/alerts/count.md)
651 - [promql/aggregate](checks/promql/aggregate.md)
652 - [query/cost](checks/query/cost.md)
653 - [rule/for](checks/rule/for.md)
654 - [rule/label](checks/rule/label.md)
655 - [rule/link](checks/rule/link.md)
656 - [rule/reject](checks/rule/reject.md)
657 Example:
658
659 ```js
660 for {
661 comment = "All alert rules must have `for: 5m` (or more) to avoid flaky alerts."
662 severity = "bug"
663 min = "5m"
664 }
665 ```
666
667## v0.55.0
668
669### Added
670
671- [labels/conflict](checks/labels/conflict.md) check will now warn if alerting
672 rule is setting labels that are already configured as external labels.
673
674### Changed
675
676- `pint watch` command now has two sub-commands: `pint watch glob` and `pint watch rule_files`.
677 See [watch mode](index.md#watch-mode) docs for details.
678- [promql/series](checks/promql/series.md) check will now ignore missing metrics if the
679 query uses a fallback with `or vector(1)`. This will reduce the number of reported
680 missing metrics for queries where the intention is to accept and ignore missing
681 metrics.
682 Example of a rule that will no longer report problems:
683
684 ```yaml
685 - alert: Foo
686 expr: sum(my_metric or vector(0)) > 1
687 ```
688
689### Fixed
690
691- Fixed false positive reports from [rule/duplicate](checks/rule/duplicate.md) when
692 using symlinks.
693- Fixed support for multi-document YAML files when using relaxed parsed mode - #746.
694
695## v0.54.0
696
697### Added
698
699- [rule/dependency](checks/rule/dependency.md) check will now warn if an alerting rule
700 that's being removed in a pull request is being used inside `ALERTS{alertname="..."}`
701 or `ALERTS_FOR_STATE{alertname="..."}` queries.
702- pint will now perform extra validation of YAML files to ensure that all values
703 are mappings and strings.
704 This will error on rules where, for example, label value is unquoted number:
705
706 Bad rule:
707
708 ```yaml
709 - alert: DeadMansSwitch
710 expr: vector(1)
711 labels:
712 priority: 5
713 ```
714
715 Good rule:
716
717 ```yaml
718 - alert: DeadMansSwitch
719 expr: vector(1)
720 labels:
721 priority: "5"
722 ```
723
724### Changed
725
726- A large part of rule parsing code was refactored and more problems will now be deduplicated.
727- Improved validation of labels and annotations.
728
729## v0.53.0
730
731### Added
732
733- pint will now report any comment that looks like a [pint control comment](./ignoring.md)
734 but cannot be parsed correctly.
735
736### Fixed
737
738- Fixed a crash when running `pint ci` and using `ignore/file` comments.
739
740## v0.52.0
741
742### Added
743
744- Both [alerts/annotation](checks/alerts/annotation.md) and [rule/label](checks/rule/label.md)
745 now support more advance validation of label and annotation values with extra `token` option.
746 In addition to the `value` regexp matching you can also validate values against a static
747 list of allowed values using new `values` option.
748 See both checks documentation for detail.
749
750### Changed
751
752- More reports will now be merged into a single comments when using BitBucket.
753
754### Fixed
755
756- Fixed YAML anchor parsing.
757- Fixed regexp matching for label names in [rule/label](checks/rule/label.md).
758
759## v0.51.1
760
761### Fixed
762
763- Fixed a few bugs in control comment parsing code.
764- BitBucket tasks were not marked as resolved correctly.
765
766## v0.51.0
767
768### Added
769
770- Added [rule/dependency](checks/rule/dependency.md) check.
771
772### Changed
773
774- When running `pint ci` pint will now first try to parse all files in current
775 working directory, before checking for files modified on current branch.
776 This is to have a full list of all rules, which is needed for checks like
777 newly added [rule/dependency](checks/rule/dependency.md).
778 This can slow `pint` runs if there's a lot of files in your repository.
779 If there are non-rule files these may fail to parse and result in check errors.
780 To avoid any errors or slowdowns from scanning unrelated files you might need
781 to add `ci` section to `.pint.hcl` with `include` and/or `exclude` options set.
782 See [examples/ci.hcl](examples/ci.hcl) for an example config.
783- `Bug` and `Fatal` severity problems are now reported as tasks when using
784 BitBucket.
785- [promql/regexp](checks/promql/regexp.md) will now check for more problems
786 with vector selectors.
787
788### Fixed
789
790- `pint` will no longer run dynamic Prometheus discovery when `--offline` flag
791 is passed.
792
793## v0.50.1
794
795### Fixed
796
797- Fixed alert preview link on `alerts/count` reports.
798
799## v0.50.0
800
801### Changed
802
803- `pint ci` now uses a new logic for deciding which rules have changed when validating
804 pull requests. Changes that would previously be invisible to pint, like modifying
805 comments or moving the entire rules file, will now trigger checks for affected
806 rules.
807- pint will now try to create fewer BitBucket comments by merging multiple
808 problem reports into a single comment.
809- Control comment handling code was refactored, there are some additional rules
810 that comment must follow. See `Control comments` section in [pint docs](./index.md).
811
812## v0.49.2
813
814### Fixed
815
816- Fixed false positive reports from `promql/regexp` - #782.
817- `Information` level reports using BitBucket were using wrong comment icon.
818
819## v0.49.1
820
821### Fixed
822
823- `alerts/count` check wasn't using `uptime` field from `prometheus` config blocks
824 for metric gap detection.
825
826## v0.49.0
827
828### Added
829
830- Added [alerts/external_labels](checks/alerts/external_labels.md) check.
831- Added support for reporting problems to TeamCity using [Service Messages](https://www.jetbrains.com/help/teamcity/service-messages.html).
832 To enable run it run `pint --teamcity lint` or `pint --teamcity ci`.
833- Problems reported to BitBucket and GitHub will now include more details.
834- Added `publicURI` field to `prometheus` configuration blocks.
835
836### Changed
837
838- When [promql/series](checks/promql/series.md) finds that a time series used
839 by a rule is missing it will now also check other defined Prometheus servers
840 and add that information to the report.
841 This allows pint to flag rules that are most likely deployed to the wrong servers,
842 using missing scrape jobs.
843- Reporting problems to BitBucket will now use comments instead of annotations.
844 This is only if there is an open pull request for tested branch, if there is no
845 open pull request problems will be reported using code insight annotations.
846
847## v0.48.2
848
849### Fixed
850
851- Fixed `pint watch` to correctly work with `discovery`.
852
853## v0.48.1
854
855- Prometheus `template` under `discovery` block can now template
856 `include` and `exclude` fields.
857
858## v0.48.0
859
860### Added
861
862- Prometheus servers can now be dynamically configured using one of
863 supported discovery mechanism: file paths or Prometheus metrics query.
864
865## v0.47.0
866
867### Added
868
869- `alerts/template` check can now report problems with alerting rules
870 when trying to use templates on a query that doesn't produce any labels
871 at all.
872 For example when using [`vector(...)`](https://prometheus.io/docs/prometheus/latest/querying/functions/#vector) function:
873
874 {% raw %}
875
876 ```yaml
877 - alert: DeadMansSwitch
878 expr: vector(1)
879 annotations:
880 summary: Deadman's switch on {{ $labels.instance }} is firing
881 ```
882
883 {% endraw %}
884
885- `alerts/comparison` check can now warn if alerting rules use a query
886 with `foo > 0 OR vector(1)`, which would always fire.
887- `alerts/template` check will now look check `on(...)` clause on binary
888 expressions. When `on(...)` is set only labels listed there will appear
889 on result metrics.
890 For example `app_type` here cannot appear on query results, even if it's
891 present on `foo` time series.
892
893 ```yaml
894 - alert: ...
895 expr: foo / on (instance, app_name) bar
896 annotations:
897 summary: ... {{ $labels.app_type }} ...
898 ```
899
900## v0.46.0
901
902### Added
903
904- Added support for `keep_firing_for` in alerting rules - #713.
905- Added `rule/keep_firing_for` check - #713.
906- Added `alerts/count` check will now estimate alerts using
907 `keep_firing_for` field if set - #713.
908- Configuration rule `match` block supports a new filter `keep_firing_for`.
909
910## v0.45.0
911
912### Added
913
914- The `query/cost` check can now use Prometheus query stats to verify query
915 evaluation time and the number of samples used by a query. See
916 [query/cost](checks/query/cost.md) docs for details.
917
918## v0.44.2
919
920### Fixed
921
922- Fixed a crash in `promql/series` check when Prometheus instance becomes
923 unavailable - #682.
924- Fixed false positive reports in `alerts/template` check - #681.
925
926## v0.44.1
927
928### Fixed
929
930- Rule names were not checked for correctly, allowing for rules with empty names
931 to pass checks.
932- Fixed GitHub annotations being added to unmodified lines - #645.
933
934## v0.44.0
935
936### Added
937
938- Added `exclude` option to `ci` config block - #609.
939- Added `minCount` & `severity` to `alerts/count` check - #612.
940 This allows to only show estimated alerts count only if there would be high
941 enough (`>= minCount`) number of alerts.
942 Setting `severity` as well allows to block rules that would create too many alerts.
943- GitHub reporter will now included folded list of all problems in the summary
944 comment - #608.
945
946### Fixed
947
948- Fixed `alerts/annotation` check regexp matching - #613.
949- When running `pint ci` using GitHub integration annotation comments are now
950 reported only on modified lines - #640.
951
952## v0.43.1
953
954### Fixed
955
956- Fixed `--base-branch` flag handling when branch name contains `/`.
957
958## v0.43.0
959
960### Added
961
962- Added `--fail-on` flag to `pint lint` command - #570.
963- Added `tls` section to `prometheus` configuration block - #540.
964
965### Changed
966
967- If a query run by pint fails because it was too expensive too run it will
968 now be reported as a warning instead of an error.
969
970### Fixed
971
972- When validating queries using `{__name__=~"...", foo="bar"}` selectors pint could end up
973 running queries matching a single label, like `count({foo="bar"})`, which could return too many
974 results. This version ensures that queries always include name matcher to avoid that.
975
976## v0.42.2
977
978### Fixed
979
980- `alerts/template` check didn't correctly handle `label_replace()` calls
981 in queries - #568.
982
983## v0.42.1
984
985### Fixed
986
987- Fixed `--base-branch` flag handling.
988 Value of this flag wasn't being used correctly - #559.
989
990## v0.42.0
991
992### Fixed
993
994- Fixed incorrect results in `promql/series` check for time series with only a single
995 data point.
996
997## v0.41.1
998
999### Fixed
1000
1001- Fixed parsing of alert `for` field values with long durations (`for: 1d`).
1002
1003## v0.41.0
1004
1005### Added
1006
1007- Added `--fail-on` flag to `pint ci` command - #525.
1008- [promql/rate](checks/promql/rate.md) will now look for queries that call
1009 `rate()` on results of `sum(counter)` via recording rules.
1010 Example:
1011
1012 ```yaml
1013 - record: my:sum
1014 expr: sum(http_requests_total)
1015
1016 - alert: my alert
1017 expr: rate(my:sum[5m])
1018 ```
1019
1020- Added [rule/for](checks/rule/for.md) check.
1021- Added `owners` configuration block for setting the list of allowed rule owner values.
1022 See [configuration](configuration.md) for details.
1023
1024### Changed
1025
1026- `pint lint` output will now include severity level as a text label - #524.
1027
1028## v0.40.1
1029
1030### Fixed
1031
1032- Fixed a bug in `pint ci` that would cause a failure if a directory was renamed.
1033- Fixed false positive reports from `promql/series` check when a time series
1034 disappears from Prometheus.
1035- Fixed Prometheus flags parsing in `promql/range_query` check.
1036
1037## v0.40.0
1038
1039### Added
1040
1041- Allow snoozing checks for entire file using `# pint file/snooze ...` comments.
1042- Added `lookbackRange` and `lookbackStep` configuration option to the
1043 [promql/series](checks/promql/series.md) check - #493.
1044
1045### Changed
1046
1047- Reverted GitHub integration to use [Pull Request Review](https://docs.github.com/en/rest/pulls/reviews)
1048 API - #490.
1049
1050## v0.39.0
1051
1052### Changed
1053
1054- GitHub integration now uses [Check Runs](https://docs.github.com/en/rest/checks/runs) API - #478.
1055
1056## v0.38.1
1057
1058### Fixed
1059
1060- `# pint file/disable` comments didn't properly handle Prometheus tags, this is fixed now.
1061
1062## v0.38.0
1063
1064### Added
1065
1066- `prometheus` configuration blocks now accepts `tags` field with a list of tags.
1067 Tags can be used to disable or snooze specific checks on all Prometheus instances
1068 with that tag.
1069 See [ignoring](ignoring.md) for details.
1070
1071## v0.37.0
1072
1073### Added
1074
1075- Added `pint_rule_file_owner` metric.
1076
1077## v0.36.0
1078
1079### Added
1080
1081- Added ability to expand environment variables in pint configuration file.
1082 See [configuration](configuration.md) for details.
1083
1084## v0.35.0
1085
1086### Added
1087
1088- Use [uber-go/automaxprocs](https://github.com/uber-go/automaxprocs) to
1089 automatically set GOMAXPROCS to match Linux container CPU quota.
1090- Added [labels/conflict](checks/labels/conflict.md) check.
1091- If you want to disable individual checks just for some time then you can now
1092 snooze them instead of disabling forever.
1093
1094 The difference between `# pint disable ...` and `# pint snooze ...` comments is that
1095 the snooze comment must include a timestamp. Selected check will be disabled _until_
1096 that timestamp.
1097 Timestamp must either use [RFC3339](https://www.rfc-editor.org/rfc/rfc3339) syntax
1098 or `YYYY-MM-DD` (if you don't care about time and want to snooze until given date).
1099 Examples:
1100
1101 ```yaml
1102 # pint snooze 2023-01-12T10:00:00Z promql/series
1103 # pint snooze 2023-01-12 promql/rate
1104 - record: ...
1105 expr: ...
1106 ```
1107
1108### Changed
1109
1110- Removed `cache` option from `prometheus` config blocks. Query cache will now auto-size itself
1111 as needed.
1112
1113 If you have a config entry with `cache` option, example:
1114
1115 ```javascript
1116 prometheus "prod" {
1117 uri = "https://prometheus.example.com"
1118 cache = 20000
1119 }
1120 ```
1121
1122 Then pint will fail to start. To fix this simply remove the `cache` option:
1123
1124 ```javascript
1125 prometheus "prod" {
1126 uri = "https://prometheus.example.com"
1127 }
1128 ```
1129
1130## v0.34.0
1131
1132### Added
1133
1134- Added [rule/duplicate](checks/rule/duplicate.md) check.
1135
1136## v0.33.1
1137
1138### Fixed
1139
1140- Fixed a regression causing poor query cache hit rate.
1141
1142## v0.33.0
1143
1144### Added
1145
1146- Added `uptime` field in `prometheus` configuration block.
1147 This field can be used to set a custom metric used for Prometheus uptime checks
1148 and by default uses `up` metric.
1149 If you have a Prometheus with a large number of scrape targets there might
1150 be a huge number of `up` time series making those uptime checks slow to run.
1151 If your Prometheus is configured to scrape itself, then you most likely want to use
1152 one of metrics exported by Prometheus, like `prometheus_build_info`:
1153
1154 ```javascript
1155 prometheus "prod" {
1156 uri = "https://prometheus.example.com"
1157 uptime = "prometheus_build_info"
1158 }
1159 ```
1160
1161### Changed
1162
1163- Refactored some queries used by [promql/series](checks/promql/series.md) check to
1164 avoid sending queries that might be very slow and/or return a huge amount of data.
1165- Prometheus query cache now takes into account the size of cached response.
1166 This makes memory usage needed for query cache more predictable.
1167 As a result the `cache` option for `prometheus` config block now means
1168 `the number of time series cached` instead of `the number of responses cached`
1169 and the default for this option is now `50000`.
1170
1171## v0.32.1
1172
1173### Fixed
1174
1175- [promql/vector_matching](checks/promql/vector_matching.md) was sending expensive
1176 queries resulting in high memory usage, this is now fixed.
1177
1178## v0.32.0
1179
1180### Added
1181
1182- Added `pint_prometheus_cache_evictions_total` metric tracking the number of times
1183 cache results were evicted from query cache.
1184- Allow disabling individual checks for the entire file using
1185 `# pint file/disable ...` comments.
1186
1187### Changed
1188
1189- Refactored query cache to only store queries that are requested more than once.
1190 This will avoid storing big responses that are never requested from the cache.
1191
1192### Fixed
1193
1194- Config validation will now check for duplicated `prometheus` block names.
1195
1196## v0.31.1
1197
1198### Fixed
1199
1200- Fixed performance regression slowing down `pint watch` over time.
1201
1202## v0.31.0
1203
1204### Added
1205
1206- `prometheus` configuration block now accepts optional `headers` field, for setting
1207 request headers that will be attached to any request made to given Prometheus server.
1208 Example:
1209
1210 ```javascript
1211 prometheus "protected" {
1212 uri = "https://prod.example.com"
1213 headers = {
1214 "X-Auth": "secret",
1215 "X-User": "bob"
1216 }
1217 ```
1218
1219### Changed
1220
1221- Prometheus range query handling was rewritten to improve memory usage
1222 caused by queries returning huge number of results.
1223 As a result pint should use up to 5x less memory.
1224
1225### Fixed
1226
1227- Fixed false positive reports in [promql/vector_matching](checks/promql/vector_matching.md)
1228 for rules using `on(...)`. Example:
1229
1230 ```yaml
1231 sum(foo) without(instance) * on(app_name) group_left() bar
1232 ```
1233
1234- Don't log passwords when Prometheus URI is using basic authentication.
1235- Fixed false positive reports in [alerts/template](checks/alerts/template.md)
1236 suggesting to use `humanize` on queries that already use `round()`.
1237- Fixed false positive reports in [alerts/comparison](checks/alerts/comparison.md)
1238 when `bool` modifier is used on a condition that is guarded by another condition.
1239 Example:
1240
1241 ```yaml
1242 alert: Foo
1243 expr: (foo > 1) > bool 1
1244 ```
1245
1246- Fixed false positive reports in [alerts/template](checks/alerts/template.md)
1247 warning about labels removed in a query despite being re-added by a join.
1248
1249## v0.30.2
1250
1251### Fixed
1252
1253- Fixed incorrect line number reporting on BitBucket annotations.
1254
1255## v0.30.1
1256
1257### Fixed
1258
1259- Fixed handling of symlinks when running `pint lint` and `pint watch` commands.
1260
1261## v0.30.0
1262
1263### Added
1264
1265- BitBucket only allows for annotations on modified lines, so when a high severity problem
1266 is reported on unmodified line pint will move that annotation to the first modified line,
1267 so it's still visible in BitBucket.
1268 Now pint will also add a note to that annotation to make it clear that the problem is really
1269 on a different line.
1270- [alerts/template](checks/alerts/template.md) will now run extra checks to validate syntax
1271 of queries executed from within alerting rule templates.
1272
1273 Example template using `sum(xxx` query that's missing closing `)`:
1274
1275 {% raw %}
1276
1277 ```yaml
1278 - alert: ...
1279 expr: ...
1280 annotations:
1281 summary: |
1282 {{ with query "sum(xxx" }}
1283 {{ . | first | value | humanize }}
1284 {{ end }}
1285 ```
1286
1287 {% endraw %}
1288
1289- If a file is ignored pint will now note that using `Information` level annotation.
1290 This will make it more obvious that a CI check passed because pint didn't run any
1291 checks due to file being excluded.
1292
1293### Changed
1294
1295- Prometheus rule files can be symlinked between directories.
1296 If the symlink source and target files are in a different directory they can
1297 end up querying different Prometheus server when running ping checks.
1298 This means that when modifying symlink target file checks must be executed against
1299 both symlink source and target.
1300 Until now pint was ignoring symlinks but starting with this release it will try to
1301 follow them. This means that if you modify a file that has symlinks pointing to them
1302 pint will try run checks against those symlinks too.
1303
1304 **NOTE**: pint can only detect and check symlinks if they are located in the current
1305 working directory (as seen by running pint process) or its sub-directories.
1306
1307### Fixed
1308
1309- Fixed a regression in [promql/vector_matching](checks/promql/vector_matching.md) that
1310 would cause a panic when parsing function calls with optional arguments.
1311
1312## v0.29.4
1313
1314### Fixed
1315
1316- [promql/vector_matching](checks/promql/vector_matching.md) was incorrectly handling
1317 queries containing function calls with multiple arguments.
1318
1319## v0.29.3
1320
1321### Fixed
1322
1323- Revert 'Use smaller buffers when decoding Prometheus API responses' change.
1324
1325## v0.29.2
1326
1327### Fixed
1328
1329- Use smaller buffers when decoding Prometheus API responses.
1330
1331## v0.29.1
1332
1333### Fixed
1334
1335- Fixed wrong request formatting for Prometheus metric metadata queries.
1336
1337## v0.29.0
1338
1339### Changed
1340
1341- Switched from using [prometheus/client_golang](https://github.com/prometheus/client_golang)
1342 API client to streaming JSON library [prymitive/current](https://github.com/prymitive/current)
1343
1344### Fixed
1345
1346- Avoid reporting same issue multiple times in `promql/rate` and `promql/regexp` checks.
1347
1348## v0.28.7
1349
1350### Changed
1351
1352- Updated Prometheus modules to [v2.38.0](https://github.com/prometheus/prometheus/releases/tag/v2.38.0).
1353 This adds support for `toTime` template function.
1354
1355## v0.28.6
1356
1357### Fixed
1358
1359- Fixed symlink handling when running `pint lint`.
1360
1361## v0.28.5
1362
1363### Fixed
1364
1365- Remove noisy debug logs.
1366
1367## v0.28.4
1368
1369### Added
1370
1371- Added `pint_prometheus_cache_miss_total` metric.
1372
1373### Changed
1374
1375- Reduce log level for `File parsed` messages.
1376
1377### Fixed
1378
1379- Purge expired cache entries faster to reduce memory usage.
1380
1381## v0.28.3
1382
1383### Fixed
1384
1385- Fix `absent()` handling in [alerts/comparison](checks/alerts/comparison.md) #330.
1386
1387## v0.28.2
1388
1389### Added
1390
1391- Added `--min-severity` flag to the `pint lint` command. Default value is set to `warning`.
1392
1393### Fixed
1394
1395- Fix a regression in [promql/vector_matching](checks/promql/vector_matching.md) introduced
1396 in previous release.
1397- Fix [promql/series](checks/promql/series.md) disable comments not working when there
1398 are multiple comments on a rule.
1399- [promql/series](checks/promql/series.md) no longer emits an information message
1400 `metric is generated by alerts ...`.
1401
1402## v0.28.1
1403
1404### Fixed
1405
1406- Don't use `topk` in [promql/vector_matching](checks/promql/vector_matching.md) check to
1407 avoid false positives.
1408
1409## v0.28.0
1410
1411### Added
1412
1413- [promql/rate](checks/promql/rate.md) check will now also validate `deriv` function usage.
1414- [alerts/annotation](checks/alerts/annotation.md) check will now recommend using one of
1415 humanize functions if alert query is returning results based on `rate()` and the value
1416 is used in annotations.
1417
1418### Changed
1419
1420- [promql/series](checks/promql/series.md) check now supports more flexible
1421 `# pint disable promql/series(...)` comments.
1422 Adding a comment `# pint disable promql/series({cluster="dev"})` will disable this check
1423 for any metric selector with `cluster="dev"` matcher.
1424- [query/cost](checks/query/cost.md) check will now calculate how much Prometheus memory
1425 will be needed for storing results of given query.
1426 `bytesPerSample` option that was previously used to calculate this was removed.
1427- `prometheus {}` config block now allows to pass a list of paths to explicitly ignore
1428 by setting `exclude` option. Existing `paths` option was renamed to `include` for
1429 consistency. Example migration:
1430
1431 ```javascript
1432 prometheus "foo" {
1433 [...]
1434 paths = [ "rules/.*" ]
1435 }
1436 ```
1437
1438 becomes
1439
1440 ```javascript
1441 prometheus "foo" {
1442 [...]
1443 include = [ "rules/.*" ]
1444 }
1445 ```
1446
1447### Fixed
1448
1449- `pint_last_run_checks` and `pint_last_run_checks_done` were not updated properly.
1450
1451## v0.27.0
1452
1453### Added
1454
1455- Deduplicate reports where possible to avoid showing same issue twice.
1456- [rule/link](checks/rule/link.md) check for validating URIs found in alerting rule annotations.
1457
1458### Changed
1459
1460- Add more details to BitBucket CI reports.
1461- More compact console output when running `pint lint`.
1462
1463## v0.26.0
1464
1465### Added
1466
1467- [promql/range_query](checks/promql/range_query.md) check.
1468
1469### Fixed
1470
1471- Strict parsing mode shouldn't fail on template errors, those will be later
1472 reported by `alerts/template` check.
1473
1474## v0.25.0
1475
1476### Changed
1477
1478- All timeout options are now optional. This includes following config blocks:
1479 - `prometheus { timeout = ... }`
1480 - `repository { bitbucket { timeout = ... } }`
1481 - `repository { github { timeout = ... } }`
1482- `pint` will now try to discover all repository settings from environment variables
1483 when run as part of GitHub Actions workflow and so it doesn't need any
1484 `repository { github { ... } }` configuration block for that anymore.
1485 Setting `GITHUB_AUTH_TOKEN` is the only requirement for GitHub Actions now.
1486
1487## v0.24.1
1488
1489### Fixed
1490
1491- Fixed line reporting on some strict parser errors.
1492
1493### Added
1494
1495- Added `--base-branch` flag to `pint ci` command.
1496
1497## v0.24.0
1498
1499### Added
1500
1501- Added rate limit for Prometheus API requests with a default value of 100
1502 requests per second. To customise it set `rateLimit` field inside selected
1503 `prometheus` server definition.
1504- Added `pint_last_run_checks` and `pint_last_run_checks_done` metrics to track
1505 progress when running `pint watch`.
1506
1507## v0.23.0
1508
1509### Fixed
1510
1511- Improved range query cache efficiency.
1512
1513### Added
1514
1515- Added extra global configuration for `promql/series` check.
1516 See check [documentation](checks/promql/series.md) for details.
1517- `prometheus` server definition in `pint` config file can now accept optional
1518 `cache` field (defaults to 10000) to allow fine tuning of built-in Prometheus
1519 API query caching.
1520- Added `pint_prometheus_cache_size` metric that exposes the number of entries
1521 currently in the query cache.
1522
1523## v0.22.2
1524
1525### Fixed
1526
1527- Improved error reporting when strict mode is enabled.
1528
1529## v0.22.1
1530
1531### Fixed
1532
1533- Fixed high memory usage when running range queries against Prometheus servers.
1534
1535## v0.22.0
1536
1537### Changed
1538
1539- The way `pint` sends API requests to Prometheus was changed to improve performance.
1540
1541 First change is that each `prometheus` server definition in `pint` config file can
1542 now accept optional `concurrency` field (defaults to 16) that sets a limit on how
1543 many concurrent requests can that server receive. There is a new metric that
1544 tracks how many queries are currently being run for each Prometheus server -
1545 `pint_prometheus_queries_running`.
1546
1547 Second change is that range queries will now be split into smaller queries, so
1548 if `pint` needs to run a range query on one week of metrics, then it will break
1549 this down into multiple queries each for a two hour slot, and then merge all
1550 the results. Previously it would try to run a single query for a whole week
1551 and if that failed it would reduce time range until a query would succeed.
1552
1553### Fixed
1554
1555- Strict parsing mode didn't fully validate rule group files, this is now fixed
1556 and pint runs the same set of checks as Prometheus.
1557- Fixed `promql/series` handling of rules with `{__name__=~"foo|bar"}` queries.
1558- If Prometheus was stopped or restarted `promql/series` would occasionally
1559 report metrics as "sometimes present". This check will now try to find time
1560 ranges with no metrics in Prometheus and ignore these when checking if
1561 metrics are present.
1562
1563## v0.21.1
1564
1565### Fixed
1566
1567- `pint_prometheus_queries_total` and `pint_prometheus_cache_hits_total` metric wasn't
1568 always correctly updated.
1569- Ignore `unknown` metric types in `promql/rate`.
1570
1571## v0.21.0
1572
1573### Added
1574
1575- `promql/rate` check will now report if `rate()` or `irate()` function is being
1576 passed a non-counter metric.
1577
1578## v0.20.0
1579
1580### Fixed
1581
1582- pint will now correctly handle YAML anchors.
1583
1584## v0.19.0
1585
1586### Added
1587
1588- Parsing files in relaxed mode will now try to find rules inside multi-line strings #252.
1589 This allows direct linting of k8s manifests like the one below:
1590
1591 ```yaml
1592 ---
1593 kind: ConfigMap
1594 apiVersion: v1
1595 metadata:
1596 name: example-app-alerts
1597 labels:
1598 app: example-app
1599 data:
1600 alerts: |
1601 groups:
1602 - name: example-app-alerts
1603 rules:
1604 - alert: Example_Is_Down
1605 expr: kube_deployment_status_replicas_available{namespace="example-app"} < 1
1606 for: 5m
1607 labels:
1608 priority: "2"
1609 environment: production
1610 annotations:
1611 summary: "No replicas for Example have been running for 5 minutes"
1612 ```
1613
1614## v0.18.1
1615
1616### Fixed
1617
1618- Fixed incorrect line reported when pint fails to unmarshal YAML file.
1619
1620## v0.18.0
1621
1622### Added
1623
1624- Allow fine tuning `promql/series` check with extra control comments
1625 `# pint rule/set promql/series min-age ...` and
1626 `# pint rule/set promql/series ignore/label-value ...`
1627 See [promql/series](checks/promql/series.md) for details.
1628- `promql/regexp` will report redundant use of regex anchors.
1629
1630### Changed
1631
1632- `promql/series` will now report missing metrics only if they were last seen
1633 over 2 hours ago by default. This can be customised per rule with comments.
1634
1635## v0.17.7
1636
1637### Fixed
1638
1639- Fix problem line reporting for `rule/owner` check.
1640- Add missing `rule/owner` documentation page.
1641
1642## v0.17.6
1643
1644### Fixed
1645
1646- Fixed false positive reports from `promql/series` check when running
1647 `pint watch`.
1648
1649## v0.17.5
1650
1651### Added
1652
1653- Added `pint_last_run_duration_seconds` metric.
1654- Added `--require-owner` flag support to `pint ci` command.
1655
1656### Fixed
1657
1658- Better handling of YAML unmarshal errors.
1659
1660## v0.17.4
1661
1662### Fixed
1663
1664- Fixed false positive reports from `alerts/template` check when `absent()` is
1665 used inside a binary expression.
1666
1667## v0.17.3
1668
1669### Fixed
1670
1671- File parse errors didn't report correct line numbers when running `pint ci`.
1672
1673## v0.17.2
1674
1675### Fixed
1676
1677- File parse errors were not reported correctly when running `pint ci`.
1678
1679## v0.17.1
1680
1681### Fixed
1682
1683- Handle `504 Gateway Timeout` HTTP responses from Prometheus same as query
1684 timeouts and retry with a shorter range query.
1685
1686## v0.17.0
1687
1688### Added
1689
1690- When running `pint ci` all checks will be skipped if any commit contains
1691 `[skip ci]` or `[no ci]` string in the commit message.
1692
1693### Changed
1694
1695- By default pint will now parse all files in strict mode, where all rule files
1696 must have the exact syntax Prometheus expects:
1697
1698 ```yaml
1699 groups:
1700 - name: example
1701 rules:
1702 - record: ...
1703 expr: ...
1704 ```
1705
1706 Previous releases were only looking for individual rules so `groups` object
1707 wasn't required. Now pint will fail to read any file that doesn't follow
1708 Prometheus syntax exactly.
1709 To enable old behaviour add `parser { relaxed = ["(.+)", ...]}` option in
1710 the config file. See [Configuration](configuration.md) for details.
1711 To enable old (relaxed) behaviour for all files add:
1712
1713 ```yaml
1714 parser {
1715 relaxed = ["(.*)"]
1716 }
1717 ```
1718
1719### Fixed
1720
1721- Improved `promql/vector_matching` checks to detect more issues.
1722- Fixed reporting of problems detected on unmodified lines when running `pint ci`.
1723
1724## v0.16.1
1725
1726### Fixed
1727
1728- Fixed false positive reports from `alerts/template` check when `absent()` function
1729 is receiving labels from a binary expression.
1730
1731## v0.16.0
1732
1733### Added
1734
1735- When running `pint watch` exported metric can include `owner` label for each rule.
1736 This is useful to route alerts based on `pint_problem` metrics to the right team.
1737 To set a rule owner add a `# pint file/owner $owner` comment in a file, to set
1738 an owner for all rules in that file. You can also set an owner per rule, by adding
1739 `# pint rule/owner $owner` comment around given rule.
1740 To enforce ownership comments in all files pass `--require-owner` flag to `pint lint`.
1741
1742## v0.15.7
1743
1744### Fixed
1745
1746- `promql/series` check no longer runs duplicated checks on source metrics when
1747 a query depends on a recording rule added in the same PR.
1748
1749## v0.15.6
1750
1751### Fixed
1752
1753- `promql/series` check was reporting that a metric stopped being exported when check
1754 queries would require a few retries.
1755
1756## v0.15.5
1757
1758### Fixed
1759
1760- `promql/series` check was reporting both `Warning` and `Bug` problems for the
1761 same metric when it was using newly added recording rule.
1762
1763## v0.15.4
1764
1765### Fixed
1766
1767- Fixed false positive reports from `promql/fragile` when `foo OR bar` is used inside
1768 aggregation.
1769
1770## v0.15.3
1771
1772### Fixed
1773
1774- Use more efficient queries for `promql/series` check.
1775- Fixed YAML parsing panics detected by Go 1.18 fuzzing.
1776
1777## v0.15.2
1778
1779### Fixed
1780
1781- Improved query cache hit rate and added `pint_prometheus_cache_hits_total` metric
1782 to track the number of cache hits.
1783
1784## v0.15.1
1785
1786### Added
1787
1788- When a range query returns `query processing would load too many samples into memory`
1789 error and we retry it with smaller time range cache this information and start with
1790 that smaller time range for future calls to speed up running `pint watch`.
1791
1792## v0.15.0
1793
1794### Changed
1795
1796- Always print the number of detected problems when running `pint lint`.
1797- `promql/series` check was refactored and will now detect a range of
1798 problems. See [promql/series](checks/promql/series.md) for details.
1799- `promql/regexp` severity is now `Bug` instead of a `Warning`.
1800- `promql/rate` check will no longer produce warnings, it will only
1801 report issues that cause queries to never return anything.
1802
1803## v0.14.0
1804
1805### Added
1806
1807- Allow matching alerting rules by `for` field - #148. Example:
1808
1809 ```js
1810 rule {
1811 match {
1812 for = ">= 10m"
1813 }
1814 }
1815 ```
1816
1817- Regexp matchers used in check rules can now reference rule fields.
1818 See [Configuration](configuration.md) for details.
1819
1820### Changed
1821
1822- Added `filename` label to `pint_problem` metric - #170.
1823- Include Prometheus server URI in reported problems.
1824
1825### Fixed
1826
1827- Fixed `pint ci` handling when a file was added to git and then removed in the
1828 next commit.
1829
1830## v0.13.2
1831
1832### Fixed
1833
1834- `yaml/parse` was using incorrect line numbers for errors caused by duplicated
1835 YAML keys.
1836
1837## v0.13.1
1838
1839### Fixed
1840
1841- Don't use fail-over Prometheus servers in case of errors caused by the query
1842 itself, like `many-to-many matching not allowed`.
1843
1844## v0.13.0
1845
1846### Added
1847
1848- `yaml/parse` error will be raised if a rule file contains duplicated keys, example:
1849
1850 ```yaml
1851 - record: foo
1852 expr: sum(my_metric)
1853 expr: sum(my_metric) without(instance)
1854 ```
1855
1856### Changed
1857
1858- `prometheus` config block now allows to specify fail-over URIs using `failover` field.
1859 If fail-over URIs are set and main URI fails to respond pint will attempt to use them
1860 in the order specified until one of them works.
1861- `prometheus` config block now allows to define how upstream errors are handled using
1862 `required` field. If `required` is set to `true` any check that depends on remote
1863 Prometheus server will be reported as `bug` if it's unable to talk to it.
1864 If `required` is set to `false` pint will only emit `warning` level results.
1865 Default value for `required` is `false`. Set it to `true` if you want to hard fail
1866 in case of remote Prometheus issues. Note that setting it to `true` might block
1867 PRs when running `pint ci` until pint is able to talk to Prometheus again.
1868- Renamed `pint/parse` to `yaml/parse` and added missing documentation for it.
1869
1870## v0.12.0
1871
1872### Added
1873
1874- Added `pint_last_run_time_seconds` and `pint_rules_parsed_total` metrics when running `pint watch`.
1875
1876### Changed
1877
1878- `promql/comparison` only applies to alerts, so it was renamed to
1879 `alerts/comparison`.
1880- Online documentation hosted at [cloudflare.github.io/pint](https://cloudflare.github.io/pint/)
1881 was reworked.
1882- `alerts/count` check will now retry range queries with shorter time window
1883 on `found duplicate series for the match group ...` errors from Prometheus.
1884
1885## v0.11.1
1886
1887### Fixed
1888
1889- `pint_prometheus_queries_total` and `pint_prometheus_query_errors_total` metrics
1890 were not incremented correctly.
1891
1892## v0.11.0
1893
1894### Added
1895
1896- Added `promql/regexp` check that will warn about unnecessary regexp matchers.
1897- Added `pint_prometheus_queries_total` and `pint_prometheus_query_errors_total`
1898 metric when running `pint watch`.
1899
1900## v0.10.1
1901
1902### Fixed
1903
1904- Fixed a number of bug with `promql/vector_matching` check.
1905
1906## v0.10.0
1907
1908### Changed
1909
1910- `query/series` check was renamed to `promql/series`.
1911
1912### Fixed
1913
1914- Improved the logic of `promql/vector_matching` check.
1915
1916## v0.9.0
1917
1918### Changed
1919
1920- Removed `lines` label from `pint_problem` metric exported when running `pint watch`.
1921- Multiple `match` and `ignore` blocks can now be specified per each `rule`.
1922
1923## v0.8.2
1924
1925### Added
1926
1927- Export `pint_version` metric when running `pint watch`.
1928- Added `--min-severity` flag to `pint watch` command.
1929
1930## v0.8.1
1931
1932### Added
1933
1934- Added `--max-problems` flag to `pint watch` command.
1935
1936### Changed
1937
1938- Updated Prometheus modules to [v2.33.0](https://github.com/prometheus/prometheus/releases/tag/v2.33.0).
1939 This adds support for `stripPort` template function.
1940
1941## v0.8.0
1942
1943### Added
1944
1945- Added new `promql/fragile` check.
1946- BitBucket reports will now include a link to documentation.
1947
1948## v0.7.3
1949
1950### Added
1951
1952- `--workers` flag to control the number of worker threads for running checks.
1953
1954## v0.7.2
1955
1956### Changed
1957
1958- More aggressive range reduction for `query processing would load too many samples into memory`
1959 errors when sending range queries to Prometheus servers.
1960
1961## v0.7.1
1962
1963### Added
1964
1965- Added `command` filter to `match` / `ignore` blocks. This allows to include
1966 skip some checks when (for example) running `pint watch` but include them
1967 in `pint lint` run.
1968
1969## v0.7.0
1970
1971### Added
1972
1973- Cache each Prometheus server responses to minimise the number of API calls.
1974- `pint watch` will start a daemon that will continuously check all matching rules
1975 and expose metrics describing all discovered problems.
1976
1977### Changed
1978
1979- `alerts/annotation` and `rule/label` now include `required` flag value in
1980 `# pint disable ...` comments.
1981 Rename `# pint disable alerts/annotation($name)` to
1982 `# pint disable alerts/annotation($name:$required)` and
1983 `# pint disable rule/label($name)` to `# pint disable rule/label($name:$required)`.
1984- `--offline` and `--disabled` flags are now global, use `pint --offline lint` instead
1985 of `pint lint --offline`.
1986
1987### Fixed
1988
1989- `promql/rate`, `query/series` and `promql/vector_matching` checks were not enabled
1990 for all defined `prometheus {}` blocks unless there was at least one `rule {}` block.
1991- `annotation` based `match` blocks didn't work correctly.
1992
1993## v0.6.6
1994
1995### Fixed
1996
1997- File renames were not handled correctly when running `git ci` on branches with
1998 multiple commits.
1999
2000## v0.6.5
2001
2002### Added
2003
2004- Allow disabling `query/series` check for individual series using
2005 `# pint disable query/series(my_metric_name)` comments.
2006
2007## v0.6.4
2008
2009### Fixed
2010
2011- Fixed docker builds.
2012
2013## v0.6.3
2014
2015### Fixed
2016
2017- `aggregate` check didn't report stripping required labels on queries
2018 using aggregation with no grouping labels (`sum(foo)`).
2019- `aggregate` check didn't test for name and label matches on alert rules.
2020
2021## v0.6.2
2022
2023### Changed
2024
2025- `template` check will now include alert query line numbers when reporting issues.
2026
2027## v0.6.1
2028
2029### Fixed
2030
2031- Labels returned by `absent()` are only from equal match types (`absent(foo="bar")`,
2032 not `absent(foo=~"bar.+")` but `alerts/template` didn't test for match type when
2033 checking for labels sourced from `absent()` queries.
2034
2035## v0.6.0
2036
2037### Changed
2038
2039- `aggregate` check was refactored and uses to run a single test for both
2040 `by` and `without` conditions. As a result this check might now find issues
2041 previously undetected.
2042 Check suppression comments will need to be migrated:
2043 - `# pint disable promql/by` becomes `# pint disable promql/aggregate`
2044 - `# pint disable promql/without` becomes `# pint disable promql/aggregate`
2045 - `# pint ignore promql/by` becomes `# pint ignore promql/aggregate`
2046 - `# pint ignore promql/without` becomes `# pint ignore promql/aggregate`
2047
2048## v0.5.3
2049
2050### Fixed
2051
2052- Fixed false positive reports in `aggregate` check.
2053
2054## v0.5.2
2055
2056### Added
2057
2058- `--no-color` flag for disabling output colouring.
2059
2060### Fixed
2061
2062- Fixed duplicated warnings when multiple `rule {...}` blocks where configured.
2063
2064## v0.5.1
2065
2066### Fixed
2067
2068- Specifying multiple `# pint disable ...` comments on a single rule would only apply
2069 last comment. This now works correctly and all comments will be applied.
2070
2071## v0.5.0
2072
2073### Added
2074
2075- Added `alerts/for` check that will look for invalid `for` values in alerting rules.
2076 This check is enabled by default.
2077
2078### Changed
2079
2080- `comparison` check is now enabled by default and require no configuration.
2081 Remove `comparison{ ... }` blocks from pint config file when upgrading.
2082- `template` check is now enabled by default and require no configuration.
2083 Remove `template{ ... }` blocks from pint config file when upgrading.
2084- `rate` check is now enabled by default for all configured Prometheus servers.
2085 Remove `rate{ ... }` blocks from pint config file when upgrading.
2086- `series` check is now enabled by default for all configured Prometheus servers.
2087 Remove `series{ ... }` blocks from pint config file when upgrading.
2088- `vector_matching` check is now enabled by default for all configured Prometheus servers.
2089 Remove `vector_matching{ ... }` blocks from pint config file when upgrading.
2090
2091## v0.4.4
2092
2093### Added
2094
2095- Support `parseDuration` function in alert templates added in Prometheus 2.32.0
2096
2097## v0.4.3
2098
2099### Fixed
2100
2101- Fixed `series` check handling of queries with `{__name__="foo"}` selectors.
2102
2103## v0.4.2
2104
2105### Fixed
2106
2107- Fixed `template` check handling of `absent` calls on aggregated metrics, like
2108 `absent(sum(nonexistent{job="myjob"}))`.
2109
2110## v0.4.1
2111
2112### Added
2113
2114- `template` check will now warn if any template is referencing a label that is not passed to
2115 `absent()`.
2116 Example:
2117
2118 {% raw %}
2119
2120 ```yaml
2121 - alert: Foo
2122 expr: absent(foo{env="prod"})
2123 annotations:
2124 summary: "foo metric is missing for job {{ $labels.job }}"
2125 ```
2126
2127 {% endraw %}
2128
2129 Would generate a warning since `absent()` can only return labels that are explicitly
2130 passed to it and the above call only passes `env` label.
2131 This can be fixed by updating the query to `absent(foo{env="prod", job="bar"})`.
2132
2133## v0.4.0
2134
2135### Added
2136
2137- `comparison` check will now warn when alert query uses
2138 [bool](https://prometheus.io/docs/prometheus/latest/querying/operators/#comparison-binary-operators)
2139 modifier after condition, which can cause alert to always fire.
2140 Example:
2141
2142 ```yaml
2143 - alert: Foo
2144 expr: rate(error_count[5m]) > bool 5
2145 ```
2146
2147 Having `bool` as part of `> 5` condition means that the query will return value `1` when condition
2148 is met, and `0` when it's not. Rather than returning value of `rate(error_count[5m])` only when
2149 that value is `> 5`. Since all results of an alerting rule `expr` are considered alerts such alert
2150 rule could always fire, regardless of the value returned by `rate(error_count[5m])`.
2151
2152### Fixed
2153
2154- `comparison` check will now ignore `absent(foo)` alert queries without any condition.
2155
2156## v0.3.1
2157
2158### Added
2159
2160- `--offline` flag for `pint ci` command.
2161
2162### Fixed
2163
2164- Fixed `template` check panic when alert query had a syntax error.
2165
2166## v0.3.0
2167
2168### Added
2169
2170- `rule` block can now specify `ignore` conditions that have the same syntax as `match`
2171 but will disable `rule` for matching alerting and recording rules #48.
2172- `match` and `ignore` blocks can now filter alerting and recording rules by name.
2173 `record` will be used as name for recording rules and `alert` for alerting rules.
2174
2175## v0.2.0
2176
2177### Added
2178
2179- `--offline` flag for `pint lint` command. When passed only checks that don't send
2180 any live queries to Prometheus server will be run.
2181- `template` check will now warn if template if referencing a label that is being
2182 stripped by aggregation.
2183 Example:
2184
2185 {% raw %}
2186
2187 ```yaml
2188 - alert: Foo
2189 expr: count(up) without(instance) == 0
2190 annotations:
2191 summary: "foo is down on {{ $labels.instance }}"
2192 ```
2193
2194 {% endraw %}
2195
2196 Would generate a warning since `instance` label is being stripped by `without(instance)`.
2197
2198## v0.1.5
2199
2200### Fixed
2201
2202- Fixed file descriptor leak due to missing file `Close()` #69.
2203
2204## v0.1.4
2205
2206### Changed
2207
2208- Retry queries that error with `query processing would load too many samples into memory`
2209 using a smaller time range.
2210
2211## v0.1.3
2212
2213### Added
2214
2215- `vector_matching` check for finding queries with incorrect `on()` or `ignoring()`
2216 keywords.
2217
2218### Fixed
2219
2220- `comparison` check would trigger false positive for rules using `unless` keyword.
2221
2222## v0.1.2
2223
2224### Fixed
2225
2226- `# pint skip/line` place between `# pint skip/begin` and `# pint skip/end` lines would
2227 reset ignore rules causing lines that should be ignored to be parsed.
2228
2229## v0.1.1
2230
2231### Changed
2232
2233- `value` check was replaced by `template`, which covers the same functionality and more.
2234 See [docs](/docs/CONFIGURATION.md#template) for details.