Changelog

v0.68.0

Changed

• pint now uses Prometheus 3.0 libraries for parsing PromQL, which adds support for new query syntax that allows for dots and UTF-8 chars in metric/label names, example:

  {"status.üp"} == 0
  

v0.67.3

Fixed

• Fixed a crash when parsing vector() calls with non-number arguments.

v0.67.2

Fixed

• Improved the accuracy of alerts/template check.

v0.67.1

Fixed

• Improved message formatting in the alerts/template check.

v0.67.0

Added

• Added --checkstyle flag to pint lint & pint ci for writing XML report file in checkstyle format - #1129.
• Added --json flag to both pint lint and pint ci commands, this enables writing a JSON file with the report of all problems #606.
• promql/fragile will now warn when alerting rules are using one of the aggregation operation that can return different series on every evaluation, which can cause alert floppiness - #820.
• promql/regexp check now supports extra configuration options to disable reports on smelly selector - #1096.

• Checks can be enabled or disabled specifically for some Prometheus rules via rule {} config blocks. Adding enable or disable option with a list of checks names allows to selectively enable or disable checks only for Prometheus rules that match given rule {} definition. Enabling checks only for matching rules will only work if these checks are disabled globally via check { disabled = [] } config block. For example to disable promql/rate check for all rules except alerting rules in the rules/critical folder:

  checks {
    // This will disable promql/rate by default.
    disabled = [ "promql/rate" ]
  }
  rule {
    match {
      path = "rules/critical/.*"
      kind = "alerting"
    }
    // This will enable promql/rate only for Prometheus rules matching all our match conditions above.
    enable = [ "promql/rate" ]
  }
  

Changed

• alerts/template check was refactored and will now produce more accurate results. Messages produced by this check might include details of the PromQL query fragment causing the problem if the query is complex enough.

Fixed

• Don’t try to create GitLab comments on unmodified lines - #1147.

v0.66.1

Fixed

• Fixed message formatting in the promql/series check.

v0.66.0

Added

• Added fallbackTimeout option to the promql/series check that controls how much time pint can spend checking other Prometheus servers for missing metrics.

Fixed

• Reverted Fixed colored output on some environments - #1106 change as it was breaking GitHub report comments.
• Fixed panics in rule/duplicate check.

v0.65.3

Fixed

• Fixed GitHub actions permissions.

v0.65.2

Fixed

• Improved accuracy of the rule/duplicate check.
• Fixed GitHub reporter trying to create pull request comments to unmodified lines - #1120.
• Fixed colored output on some environments - #1106.
• Show correct line number when reporting YAML syntax errors.

v0.65.1

Fixed

• Fixed a bug in match block state handling that caused all rules to always match any state.

v0.65.0

Added

• promql/regexp check will now look for smelly regexp selectors. See check docs for details.
• promql/range_query now allows to configure a custom maximum duration for range queries - #1064.
• Added --enabled flag to the pint command. Passing this flag will only run selected check(s).
• Added state option to the rule match block. See configuration docs for details.

Fixed

• Don’t try to report problem on unmodified files when using GitHub reporter.
• If there is a pint config file present then pint will now always add it to the parser block exclude list. This is to avoid trying to parse it as a rule file if it’s included in the same folder as rules.

v0.64.1

Fixed

• Validate ignoreLabelsValue option values in the pint config.

v0.64.0

Added

• Added ignoreLabelsValue to promql/series check settings.
• Added include & exclude options were moved from the ci to the parser configuration block. They now apply to all pint commands, not just pint ci - #631.

v0.63.0

Added

• Added alerts/absent check.
• Added rule/name check - #1020.

Changed

• promql/vector_matching will now report more details, including which Prometheus server reports problems and which part of the query is the issue.
• GitHub report code was refactored, it should behave as before.
• When running pint ci pint will now run all checks on unmodified rules when a file/disable comment was removed.

Fixed

• Fixed false positive warnings from alerts/comparison when using absent_over_time().
• GitHub reporter will now wait before making more requests if it’s rate limited - #699.

v0.62.2

Fixed

• When using BitBucket reporter pint ci might create a comment longer than the limit allowed by BitBucket. To avoid this pint will now truncate long comments.

v0.62.1

Fixed

• Fixed false positive warnings from rule/dependency check.

v0.62.0

Added

• promql/series check will now generate warnings if there are # pint disable or # pint rule/set comments that are not matching any valid query selector or Prometheus server.

Fixed

• promql/series will now parse rule/set comments that target specific time series selectors in PromQL the same way as # pint disable comments do.

v0.61.2

Fixed

• Fixed false positive reports about removed rules when running pint ci on a branch that contains YAML syntax errors.

v0.61.1

Fixed

• Fixed release workflow on GitHub.

v0.61.0

Fixed

• Don’t suggest using humanize when alert template is already using printf on format the $value.
• Fixed git history parsing when running pint ci on a branch that include merge commits.

Changed

• Refactored YAML syntax checks to avoid using rulefmt.Parse and effectively parsing rules twice. Some error messages will have different formatting.

v0.60.0

Fixed

• alerts/annotations and rule/label will now report when a label or annotation is required but set to an empty value.
• Fixed handling of # pint ignore/line comments on line including multiple # characters.
• When reporting problems all messages will now use publicURI from each prometheus definition.

v0.59.0

Added

• Added GitLab support when running pint ci.

Changed

• Reduced severity of problems reported by promql/counter due to high number of false positives. It will now report warnings instead of errors.

Fixed

• Fixed false positive reports from rule/duplicate when a rule file is being deleted and the rules are moved to a different file.

v0.58.1

Fixed

• Correctly handle default Prometheus retention value in promql/range_query check - #958.

v0.58.0

Fixed

• Fixed false positives from rule/duplicate check when running pint ci on files that are both edited and renamed in the same PR.
• rule/dependency will no longer report removed symlinks.

v0.57.3

Fixed

• Fixed compatibility with older git releases.
• Fixed maxComments defaulting to zero when using GitHub reporter.
• Fixed maxComments comment creation on GitHub - #935.

v0.57.2

Fixed

• Fixed release workflow on GitHub actions.

v0.57.0

Added

• GitHub and BitBucket reporters now supports maxComments option to limit the number of comments pint can create on a single pull request. Default value of maxComments is 50.

v0.56.2

Fixed

• query/cost will now only create reports if a query is more expensive than any of the configured limits.
• Fixed duplicated reports when using BitBucket reporter.

v0.56.1

Fixed

• promql/counter check will now consider a metric to be a counter only if all metadata entries for it use TYPE counter. Previously it would check for at least one metadata entry with TYPE counter.

v0.56.0

Added

• Added promql/counter check - #655.
• Check using custom rules now accept an optional comment option for adding a text comment to all reported problems. Checks supporting comment option:
  • alerts/annotation
  • alerts/count
  • promql/aggregate
  • query/cost
  • rule/for
  • rule/label
  • rule/link
  • rule/reject Example:

  for {
    comment  = "All alert rules must have `for: 5m` (or more) to avoid flaky alerts."
    severity = "bug"
    min      = "5m"
  }
  

v0.55.0

Added

• labels/conflict check will now warn if alerting rule is setting labels that are already configured as external labels.

Changed

• pint watch command now has two sub-commands: pint watch glob and pint watch rule_files. See watch mode docs for details.

• promql/series check will now ignore missing metrics if the query uses a fallback with or vector(1). This will reduce the number of reported missing metrics for queries where the intention is to accept and ignore missing metrics. Example of a rule that will no longer report problems:

  - alert: Foo
    expr: sum(my_metric or vector(0)) > 1
  

Fixed

• Fixed false positive reports from rule/duplicate when using symlinks.
• Fixed support for multi-document YAML files when using relaxed parsed mode - #746.

v0.54.0

Added

• rule/dependency check will now warn if an alerting rule that’s being removed in a pull request is being used inside ALERTS{alertname="..."} or ALERTS_FOR_STATE{alertname="..."} queries.

• pint will now perform extra validation of YAML files to ensure that all values are mappings and strings. This will error on rules where, for example, label value is unquoted number:

  Bad rule:

  - alert: DeadMansSwitch
    expr: vector(1)
    labels:
      priority: 5
  

  Good rule:

  - alert: DeadMansSwitch
    expr: vector(1)
    labels:
      priority: "5"
  

Changed

• A large part of rule parsing code was refactored and more problems will now be deduplicated.
• Improved validation of labels and annotations.

v0.53.0

Added

• pint will now report any comment that looks like a pint control comment but cannot be parsed correctly.

Fixed

• Fixed a crash when running pint ci and using ignore/file comments.

v0.52.0

Added

• Both alerts/annotation and rule/label now support more advance validation of label and annotation values with extra token option. In addition to the value regexp matching you can also validate values against a static list of allowed values using new values option. See both checks documentation for detail.

Changed

• More reports will now be merged into a single comments when using BitBucket.

Fixed

• Fixed YAML anchor parsing.
• Fixed regexp matching for label names in rule/label.

v0.51.1

Fixed

• Fixed a few bugs in control comment parsing code.
• BitBucket tasks were not marked as resolved correctly.

v0.51.0

Added

• Added rule/dependency check.

Changed

• When running pint ci pint will now first try to parse all files in current working directory, before checking for files modified on current branch. This is to have a full list of all rules, which is needed for checks like newly added rule/dependency. This can slow pint runs if there’s a lot of files in your repository. If there are non-rule files these may fail to parse and result in check errors. To avoid any errors or slowdowns from scanning unrelated files you might need to add ci section to .pint.hcl with include and/or exclude options set. See examples/ci.hcl for an example config.
• Bug and Fatal severity problems are now reported as tasks when using BitBucket.
• promql/regexp will now check for more problems with vector selectors.

Fixed

• pint will no longer run dynamic Prometheus discovery when --offline flag is passed.

v0.50.1

Fixed

• Fixed alert preview link on alerts/count reports.

v0.50.0

Changed

• pint ci now uses a new logic for deciding which rules have changed when validating pull requests. Changes that would previously be invisible to pint, like modifying comments or moving the entire rules file, will now trigger checks for affected rules.
• pint will now try to create fewer BitBucket comments by merging multiple problem reports into a single comment.
• Control comment handling code was refactored, there are some additional rules that comment must follow. See Control comments section in pint docs.

v0.49.2

Fixed

• Fixed false positive reports from promql/regexp - #782.
• Information level reports using BitBucket were using wrong comment icon.

v0.49.1

Fixed

• alerts/count check wasn’t using uptime field from prometheus config blocks for metric gap detection.

v0.49.0

Added

• Added alerts/external_labels check.
• Added support for reporting problems to TeamCity using Service Messages. To enable run it run pint --teamcity lint or pint --teamcity ci.
• Problems reported to BitBucket and GitHub will now include more details.
• Added publicURI field to prometheus configuration blocks.

Changed

• When promql/series finds that a time series used by a rule is missing it will now also check other defined Prometheus servers and add that information to the report. This allows pint to flag rules that are most likely deployed to the wrong servers, using missing scrape jobs.
• Reporting problems to BitBucket will now use comments instead of annotations. This is only if there is an open pull request for tested branch, if there is no open pull request problems will be reported using code insight annotations.

v0.48.2

Fixed

• Fixed pint watch to correctly work with discovery.

v0.48.1

• Prometheus template under discovery block can now template include and exclude fields.

v0.48.0

Added

• Prometheus servers can now be dynamically configured using one of supported discovery mechanism: file paths or Prometheus metrics query.

v0.47.0

Added

• alerts/template check can now report problems with alerting rules when trying to use templates on a query that doesn’t produce any labels at all. For example when using vector(...) function:

  - alert: DeadMansSwitch
    expr: vector(1)
    annotations:
      summary: Deadman's switch on {{ $labels.instance }} is firing
  

• alerts/comparison check can now warn if alerting rules use a query with foo > 0 OR vector(1), which would always fire.

• alerts/template check will now look check on(...) clause on binary expressions. When on(...) is set only labels listed there will appear on result metrics. For example app_type here cannot appear on query results, even if it’s present on foo time series.

  - alert: ...
    expr: foo / on (instance, app_name) bar
    annotations:
      summary: ...  ...
  

v0.46.0

Added

• Added support for keep_firing_for in alerting rules - #713.
• Added rule/keep_firing_for check - #713.
• Added alerts/count check will now estimate alerts using keep_firing_for field if set - #713.
• Configuration rule match block supports a new filter keep_firing_for.

v0.45.0

Added

• The query/cost check can now use Prometheus query stats to verify query evaluation time and the number of samples used by a query. See query/cost docs for details.

v0.44.2

Fixed

• Fixed a crash in promql/series check when Prometheus instance becomes unavailable - #682.
• Fixed false positive reports in alerts/template check - #681.

v0.44.1

Fixed

• Rule names were not checked for correctly, allowing for rules with empty names to pass checks.
• Fixed GitHub annotations being added to unmodified lines - #645.

v0.44.0

Added

• Added exclude option to ci config block - #609.
• Added minCount & severity to alerts/count check - #612. This allows to only show estimated alerts count only if there would be high enough (>= minCount) number of alerts. Setting severity as well allows to block rules that would create too many alerts.
• GitHub reporter will now included folded list of all problems in the summary comment - #608.

Fixed

• Fixed alerts/annotation check regexp matching - #613.
• When running pint ci using GitHub integration annotation comments are now reported only on modified lines - #640.

v0.43.1

Fixed

• Fixed --base-branch flag handling when branch name contains /.

v0.43.0

Added

• Added --fail-on flag to pint lint command - #570.
• Added tls section to prometheus configuration block - #540.

Changed

• If a query run by pint fails because it was too expensive too run it will now be reported as a warning instead of an error.

Fixed

• When validating queries using {__name__=~"...", foo="bar"} selectors pint could end up running queries matching a single label, like count({foo="bar"}), which could return too many results. This version ensures that queries always include name matcher to avoid that.

v0.42.2

Fixed

• alerts/template check didn’t correctly handle label_replace() calls in queries - #568.

v0.42.1

Fixed

• Fixed --base-branch flag handling. Value of this flag wasn’t being used correctly - #559.

v0.42.0

Fixed

• Fixed incorrect results in promql/series check for time series with only a single data point.

v0.41.1

Fixed

• Fixed parsing of alert for field values with long durations (for: 1d).

v0.41.0

Added

• Added --fail-on flag to pint ci command - #525.

• promql/rate will now look for queries that call rate() on results of sum(counter) via recording rules. Example:

  - record: my:sum
    expr: sum(http_requests_total)
  
  - alert: my alert
    expr: rate(my:sum[5m])
  

• Added rule/for check.
• Added owners configuration block for setting the list of allowed rule owner values. See configuration for details.

Changed

• pint lint output will now include severity level as a text label - #524.

v0.40.1

Fixed

• Fixed a bug in pint ci that would cause a failure if a directory was renamed.
• Fixed false positive reports from promql/series check when a time series disappears from Prometheus.
• Fixed Prometheus flags parsing in promql/range_query check.

v0.40.0

Added

• Allow snoozing checks for entire file using # pint file/snooze ... comments.
• Added lookbackRange and lookbackStep configuration option to the promql/series check - #493.

Changed

• Reverted GitHub integration to use Pull Request Review API - #490.

v0.39.0

Changed

• GitHub integration now uses Check Runs API - #478.

v0.38.1

Fixed

• # pint file/disable comments didn’t properly handle Prometheus tags, this is fixed now.

v0.38.0

Added

• prometheus configuration blocks now accepts tags field with a list of tags. Tags can be used to disable or snooze specific checks on all Prometheus instances with that tag. See ignoring for details.

v0.37.0

Added

• Added pint_rule_file_owner metric.

v0.36.0

Added

• Added ability to expand environment variables in pint configuration file. See configuration for details.

v0.35.0

Added

• Use uber-go/automaxprocs to automatically set GOMAXPROCS to match Linux container CPU quota.
• Added labels/conflict check.

• If you want to disable individual checks just for some time then you can now snooze them instead of disabling forever.

  The difference between # pint disable ... and # pint snooze ... comments is that the snooze comment must include a timestamp. Selected check will be disabled until that timestamp. Timestamp must either use RFC3339 syntax or YYYY-MM-DD (if you don’t care about time and want to snooze until given date). Examples:

  # pint snooze 2023-01-12T10:00:00Z promql/series
  # pint snooze 2023-01-12 promql/rate
  - record: ...
    expr: ...
  

Changed

• Removed cache option from prometheus config blocks. Query cache will now auto-size itself as needed.

  If you have a config entry with cache option, example:

  prometheus "prod" {
    uri   = "https://prometheus.example.com"
    cache = 20000
  }
  

  Then pint will fail to start. To fix this simply remove the cache option:

  prometheus "prod" {
    uri = "https://prometheus.example.com"
  }
  

v0.34.0

Added

• Added rule/duplicate check.

v0.33.1

Fixed

• Fixed a regression causing poor query cache hit rate.

v0.33.0

Added

• Added uptime field in prometheus configuration block. This field can be used to set a custom metric used for Prometheus uptime checks and by default uses up metric. If you have a Prometheus with a large number of scrape targets there might be a huge number of up time series making those uptime checks slow to run. If your Prometheus is configured to scrape itself, then you most likely want to use one of metrics exported by Prometheus, like prometheus_build_info:

  prometheus "prod" {
    uri    = "https://prometheus.example.com"
    uptime = "prometheus_build_info"
  }
  

Changed

• Refactored some queries used by promql/series check to avoid sending queries that might be very slow and/or return a huge amount of data.
• Prometheus query cache now takes into account the size of cached response. This makes memory usage needed for query cache more predictable. As a result the cache option for prometheus config block now means the number of time series cached instead of the number of responses cached and the default for this option is now 50000.

v0.32.1

Fixed

• promql/vector_matching was sending expensive queries resulting in high memory usage, this is now fixed.

v0.32.0

Added

• Added pint_prometheus_cache_evictions_total metric tracking the number of times cache results were evicted from query cache.
• Allow disabling individual checks for the entire file using # pint file/disable ... comments.

Changed

• Refactored query cache to only store queries that are requested more than once. This will avoid storing big responses that are never requested from the cache.

Fixed

• Config validation will now check for duplicated prometheus block names.

v0.31.1

Fixed

• Fixed performance regression slowing down pint watch over time.

v0.31.0

Added

• prometheus configuration block now accepts optional headers field, for setting request headers that will be attached to any request made to given Prometheus server. Example:

  prometheus "protected" {
  uri      = "https://prod.example.com"
  headers  = {
    "X-Auth": "secret",
    "X-User": "bob"
  }
  

Changed

• Prometheus range query handling was rewritten to improve memory usage caused by queries returning huge number of results. As a result pint should use up to 5x less memory.

Fixed

• Fixed false positive reports in promql/vector_matching for rules using on(...). Example:

  sum(foo) without(instance) * on(app_name) group_left() bar
  

• Don’t log passwords when Prometheus URI is using basic authentication.
• Fixed false positive reports in alerts/template suggesting to use humanize on queries that already use round().

• Fixed false positive reports in alerts/comparison when bool modifier is used on a condition that is guarded by another condition. Example:

  alert: Foo
  expr: (foo > 1) > bool 1
  

• Fixed false positive reports in alerts/template warning about labels removed in a query despite being re-added by a join.

v0.30.2

Fixed

• Fixed incorrect line number reporting on BitBucket annotations.

v0.30.1

Fixed

• Fixed handling of symlinks when running pint lint and pint watch commands.

v0.30.0

Added

• BitBucket only allows for annotations on modified lines, so when a high severity problem is reported on unmodified line pint will move that annotation to the first modified line, so it’s still visible in BitBucket. Now pint will also add a note to that annotation to make it clear that the problem is really on a different line.

• alerts/template will now run extra checks to validate syntax of queries executed from within alerting rule templates.

  Example template using sum(xxx query that’s missing closing ):

  - alert: ...
    expr: ...
    annotations:
      summary: |
        {{ with query "sum(xxx" }}
        {{ . | first | value | humanize }}
        {{ end }}
  

• If a file is ignored pint will now note that using Information level annotation. This will make it more obvious that a CI check passed because pint didn’t run any checks due to file being excluded.

Changed

• Prometheus rule files can be symlinked between directories. If the symlink source and target files are in a different directory they can end up querying different Prometheus server when running ping checks. This means that when modifying symlink target file checks must be executed against both symlink source and target. Until now pint was ignoring symlinks but starting with this release it will try to follow them. This means that if you modify a file that has symlinks pointing to them pint will try run checks against those symlinks too.

  NOTE: pint can only detect and check symlinks if they are located in the current working directory (as seen by running pint process) or its sub-directories.

Fixed

• Fixed a regression in promql/vector_matching that would cause a panic when parsing function calls with optional arguments.

v0.29.4

Fixed

• promql/vector_matching was incorrectly handling queries containing function calls with multiple arguments.

v0.29.3

Fixed

• Revert ‘Use smaller buffers when decoding Prometheus API responses’ change.

v0.29.2

Fixed

• Use smaller buffers when decoding Prometheus API responses.

v0.29.1

Fixed

• Fixed wrong request formatting for Prometheus metric metadata queries.

v0.29.0

Changed

• Switched from using prometheus/client_golang API client to streaming JSON library prymitive/current

Fixed

• Avoid reporting same issue multiple times in promql/rate and promql/regexp checks.

v0.28.7

Changed

• Updated Prometheus modules to v2.38.0. This adds support for toTime template function.

v0.28.6

Fixed

• Fixed symlink handling when running pint lint.

v0.28.5

Fixed

• Remove noisy debug logs.

v0.28.4

Added

• Added pint_prometheus_cache_miss_total metric.

Changed

• Reduce log level for File parsed messages.

Fixed

• Purge expired cache entries faster to reduce memory usage.

v0.28.3

Fixed

• Fix absent() handling in alerts/comparison #330.

v0.28.2

Added

• Added --min-severity flag to the pint lint command. Default value is set to warning.

Fixed

• Fix a regression in promql/vector_matching introduced in previous release.
• Fix promql/series disable comments not working when there are multiple comments on a rule.
• promql/series no longer emits an information message metric is generated by alerts ....

v0.28.1

Fixed

• Don’t use topk in promql/vector_matching check to avoid false positives.

v0.28.0

Added

• promql/rate check will now also validate deriv function usage.
• alerts/annotation check will now recommend using one of humanize functions if alert query is returning results based on rate() and the value is used in annotations.

Changed

• promql/series check now supports more flexible # pint disable promql/series(...) comments. Adding a comment # pint disable promql/series({cluster="dev"}) will disable this check for any metric selector with cluster="dev" matcher.
• query/cost check will now calculate how much Prometheus memory will be needed for storing results of given query. bytesPerSample option that was previously used to calculate this was removed.

• prometheus {} config block now allows to pass a list of paths to explicitly ignore by setting exclude option. Existing paths option was renamed to include for consistency. Example migration:

  prometheus "foo" {
    [...]
    paths = [ "rules/.*" ]
  }
  

  becomes

  prometheus "foo" {
    [...]
    include = [ "rules/.*" ]
  }
  

Fixed

• pint_last_run_checks and pint_last_run_checks_done were not updated properly.

v0.27.0

Added

• Deduplicate reports where possible to avoid showing same issue twice.
• rule/link check for validating URIs found in alerting rule annotations.

Changed

• Add more details to BitBucket CI reports.
• More compact console output when running pint lint.

v0.26.0

Added

• promql/range_query check.

Fixed

• Strict parsing mode shouldn’t fail on template errors, those will be later reported by alerts/template check.

v0.25.0

Changed

• All timeout options are now optional. This includes following config blocks:
  • prometheus { timeout = ... }
  • repository { bitbucket { timeout = ... } }
  • repository { github { timeout = ... } }
• pint will now try to discover all repository settings from environment variables when run as part of GitHub Actions workflow and so it doesn’t need any repository { github { ... } } configuration block for that anymore. Setting GITHUB_AUTH_TOKEN is the only requirement for GitHub Actions now.

v0.24.1

Fixed

• Fixed line reporting on some strict parser errors.

Added

• Added --base-branch flag to pint ci command.

v0.24.0

Added

• Added rate limit for Prometheus API requests with a default value of 100 requests per second. To customise it set rateLimit field inside selected prometheus server definition.
• Added pint_last_run_checks and pint_last_run_checks_done metrics to track progress when running pint watch.

v0.23.0

Fixed

• Improved range query cache efficiency.

Added

• Added extra global configuration for promql/series check. See check documentation for details.
• prometheus server definition in pint config file can now accept optional cache field (defaults to 10000) to allow fine tuning of built-in Prometheus API query caching.
• Added pint_prometheus_cache_size metric that exposes the number of entries currently in the query cache.

v0.22.2

Fixed

• Improved error reporting when strict mode is enabled.

v0.22.1

Fixed

• Fixed high memory usage when running range queries against Prometheus servers.

v0.22.0

Changed

• The way pint sends API requests to Prometheus was changed to improve performance.

  First change is that each prometheus server definition in pint config file can now accept optional concurrency field (defaults to 16) that sets a limit on how many concurrent requests can that server receive. There is a new metric that tracks how many queries are currently being run for each Prometheus server - pint_prometheus_queries_running.

  Second change is that range queries will now be split into smaller queries, so if pint needs to run a range query on one week of metrics, then it will break this down into multiple queries each for a two hour slot, and then merge all the results. Previously it would try to run a single query for a whole week and if that failed it would reduce time range until a query would succeed.

Fixed

• Strict parsing mode didn’t fully validate rule group files, this is now fixed and pint runs the same set of checks as Prometheus.
• Fixed promql/series handling of rules with {__name__=~"foo|bar"} queries.
• If Prometheus was stopped or restarted promql/series would occasionally report metrics as “sometimes present”. This check will now try to find time ranges with no metrics in Prometheus and ignore these when checking if metrics are present.

v0.21.1

Fixed

• pint_prometheus_queries_total and pint_prometheus_cache_hits_total metric wasn’t always correctly updated.
• Ignore unknown metric types in promql/rate.

v0.21.0

Added

• promql/rate check will now report if rate() or irate() function is being passed a non-counter metric.

v0.20.0

Fixed

• pint will now correctly handle YAML anchors.

v0.19.0

Added

• Parsing files in relaxed mode will now try to find rules inside multi-line strings #252. This allows direct linting of k8s manifests like the one below:

  ---
  kind: ConfigMap
  apiVersion: v1
  metadata:
    name: example-app-alerts
    labels:
    app: example-app
  data:
    alerts: |
      groups:
        - name: example-app-alerts
          rules:
            - alert: Example_Is_Down
              expr: kube_deployment_status_replicas_available{namespace="example-app"} < 1
              for: 5m
              labels:
                priority: "2"
                environment: production
              annotations:
                summary: "No replicas for Example have been running for 5 minutes"
  

v0.18.1

Fixed

• Fixed incorrect line reported when pint fails to unmarshal YAML file.

v0.18.0

Added

• Allow fine tuning promql/series check with extra control comments # pint rule/set promql/series min-age ... and # pint rule/set promql/series ignore/label-value ... See promql/series for details.
• promql/regexp will report redundant use of regex anchors.

Changed

• promql/series will now report missing metrics only if they were last seen over 2 hours ago by default. This can be customised per rule with comments.

v0.17.7

Fixed

• Fix problem line reporting for rule/owner check.
• Add missing rule/owner documentation page.

v0.17.6

Fixed

• Fixed false positive reports from promql/series check when running pint watch.

v0.17.5

Added

• Added pint_last_run_duration_seconds metric.
• Added --require-owner flag support to pint ci command.

Fixed

• Better handling of YAML unmarshal errors.

v0.17.4

Fixed

• Fixed false positive reports from alerts/template check when absent() is used inside a binary expression.

v0.17.3

Fixed

• File parse errors didn’t report correct line numbers when running pint ci.

v0.17.2

Fixed

• File parse errors were not reported correctly when running pint ci.

v0.17.1

Fixed

• Handle 504 Gateway Timeout HTTP responses from Prometheus same as query timeouts and retry with a shorter range query.

v0.17.0

Added

• When running pint ci all checks will be skipped if any commit contains [skip ci] or [no ci] string in the commit message.

Changed

• By default pint will now parse all files in strict mode, where all rule files must have the exact syntax Prometheus expects:

  groups:
    - name: example
      rules:
        - record: ...
          expr: ...
  

  Previous releases were only looking for individual rules so groups object wasn’t required. Now pint will fail to read any file that doesn’t follow Prometheus syntax exactly. To enable old behaviour add parser { relaxed = ["(.+)", ...]} option in the config file. See Configuration for details. To enable old (relaxed) behaviour for all files add:

  parser {
  relaxed = ["(.*)"]
  }
  

Fixed

• Improved promql/vector_matching checks to detect more issues.
• Fixed reporting of problems detected on unmodified lines when running pint ci.

v0.16.1

Fixed

• Fixed false positive reports from alerts/template check when absent() function is receiving labels from a binary expression.

v0.16.0

Added

• When running pint watch exported metric can include owner label for each rule. This is useful to route alerts based on pint_problem metrics to the right team. To set a rule owner add a # pint file/owner $owner comment in a file, to set an owner for all rules in that file. You can also set an owner per rule, by adding # pint rule/owner $owner comment around given rule. To enforce ownership comments in all files pass --require-owner flag to pint lint.

v0.15.7

Fixed

• promql/series check no longer runs duplicated checks on source metrics when a query depends on a recording rule added in the same PR.

v0.15.6

Fixed

• promql/series check was reporting that a metric stopped being exported when check queries would require a few retries.

v0.15.5

Fixed

• promql/series check was reporting both Warning and Bug problems for the same metric when it was using newly added recording rule.

v0.15.4

Fixed

• Fixed false positive reports from promql/fragile when foo OR bar is used inside aggregation.

v0.15.3

Fixed

• Use more efficient queries for promql/series check.
• Fixed YAML parsing panics detected by Go 1.18 fuzzing.

v0.15.2

Fixed

• Improved query cache hit rate and added pint_prometheus_cache_hits_total metric to track the number of cache hits.

v0.15.1

Added

• When a range query returns query processing would load too many samples into memory error and we retry it with smaller time range cache this information and start with that smaller time range for future calls to speed up running pint watch.

v0.15.0

Changed

• Always print the number of detected problems when running pint lint.
• promql/series check was refactored and will now detect a range of problems. See promql/series for details.
• promql/regexp severity is now Bug instead of a Warning.
• promql/rate check will no longer produce warnings, it will only report issues that cause queries to never return anything.

v0.14.0

Added

• Allow matching alerting rules by for field - #148. Example:

  rule {
    match {
      for = ">= 10m"
    }
  }
  

• Regexp matchers used in check rules can now reference rule fields. See Configuration for details.

Changed

• Added filename label to pint_problem metric - #170.
• Include Prometheus server URI in reported problems.

Fixed

• Fixed pint ci handling when a file was added to git and then removed in the next commit.

v0.13.2

Fixed

• yaml/parse was using incorrect line numbers for errors caused by duplicated YAML keys.

v0.13.1

Fixed

• Don’t use fail-over Prometheus servers in case of errors caused by the query itself, like many-to-many matching not allowed.

v0.13.0

Added

• yaml/parse error will be raised if a rule file contains duplicated keys, example:

  - record: foo
    expr: sum(my_metric)
    expr: sum(my_metric) without(instance)
  

Changed

• prometheus config block now allows to specify fail-over URIs using failover field. If fail-over URIs are set and main URI fails to respond pint will attempt to use them in the order specified until one of them works.
• prometheus config block now allows to define how upstream errors are handled using required field. If required is set to true any check that depends on remote Prometheus server will be reported as bug if it’s unable to talk to it. If required is set to false pint will only emit warning level results. Default value for required is false. Set it to true if you want to hard fail in case of remote Prometheus issues. Note that setting it to true might block PRs when running pint ci until pint is able to talk to Prometheus again.
• Renamed pint/parse to yaml/parse and added missing documentation for it.

v0.12.0

Added

• Added pint_last_run_time_seconds and pint_rules_parsed_total metrics when running pint watch.

Changed

• promql/comparison only applies to alerts, so it was renamed to alerts/comparison.
• Online documentation hosted at cloudflare.github.io/pint was reworked.
• alerts/count check will now retry range queries with shorter time window on found duplicate series for the match group ... errors from Prometheus.

v0.11.1

Fixed

• pint_prometheus_queries_total and pint_prometheus_query_errors_total metrics were not incremented correctly.

v0.11.0

Added

• Added promql/regexp check that will warn about unnecessary regexp matchers.
• Added pint_prometheus_queries_total and pint_prometheus_query_errors_total metric when running pint watch.

v0.10.1

Fixed

• Fixed a number of bug with promql/vector_matching check.

v0.10.0

Changed

• query/series check was renamed to promql/series.

Fixed

• Improved the logic of promql/vector_matching check.

v0.9.0

Changed

• Removed lines label from pint_problem metric exported when running pint watch.
• Multiple match and ignore blocks can now be specified per each rule.

v0.8.2

Added

• Export pint_version metric when running pint watch.
• Added --min-severity flag to pint watch command.

v0.8.1

Added

• Added --max-problems flag to pint watch command.

Changed

• Updated Prometheus modules to v2.33.0. This adds support for stripPort template function.

v0.8.0

Added

• Added new promql/fragile check.
• BitBucket reports will now include a link to documentation.

v0.7.3

Added

• --workers flag to control the number of worker threads for running checks.

v0.7.2

Changed

• More aggressive range reduction for query processing would load too many samples into memory errors when sending range queries to Prometheus servers.

v0.7.1

Added

• Added command filter to match / ignore blocks. This allows to include skip some checks when (for example) running pint watch but include them in pint lint run.

v0.7.0

Added

• Cache each Prometheus server responses to minimise the number of API calls.
• pint watch will start a daemon that will continuously check all matching rules and expose metrics describing all discovered problems.

Changed

• alerts/annotation and rule/label now include required flag value in # pint disable ... comments. Rename # pint disable alerts/annotation($name) to # pint disable alerts/annotation($name:$required) and # pint disable rule/label($name) to # pint disable rule/label($name:$required).
• --offline and --disabled flags are now global, use pint --offline lint instead of pint lint --offline.

Fixed

• promql/rate, query/series and promql/vector_matching checks were not enabled for all defined prometheus {} blocks unless there was at least one rule {} block.
• annotation based match blocks didn’t work correctly.

v0.6.6

Fixed

• File renames were not handled correctly when running git ci on branches with multiple commits.

v0.6.5

Added

• Allow disabling query/series check for individual series using # pint disable query/series(my_metric_name) comments.

v0.6.4

Fixed

• Fixed docker builds.

v0.6.3

Fixed

• aggregate check didn’t report stripping required labels on queries using aggregation with no grouping labels (sum(foo)).
• aggregate check didn’t test for name and label matches on alert rules.

v0.6.2

Changed

• template check will now include alert query line numbers when reporting issues.

v0.6.1

Fixed

• Labels returned by absent() are only from equal match types (absent(foo="bar"), not absent(foo=~"bar.+") but alerts/template didn’t test for match type when checking for labels sourced from absent() queries.

v0.6.0

Changed

• aggregate check was refactored and uses to run a single test for both by and without conditions. As a result this check might now find issues previously undetected. Check suppression comments will need to be migrated:
  • # pint disable promql/by becomes # pint disable promql/aggregate
  • # pint disable promql/without becomes # pint disable promql/aggregate
  • # pint ignore promql/by becomes # pint ignore promql/aggregate
  • # pint ignore promql/without becomes # pint ignore promql/aggregate

v0.5.3

Fixed

• Fixed false positive reports in aggregate check.

v0.5.2

Added

• --no-color flag for disabling output colouring.

Fixed

• Fixed duplicated warnings when multiple rule {...} blocks where configured.

v0.5.1

Fixed

• Specifying multiple # pint disable ... comments on a single rule would only apply last comment. This now works correctly and all comments will be applied.

v0.5.0

Added

• Added alerts/for check that will look for invalid for values in alerting rules. This check is enabled by default.

Changed

• comparison check is now enabled by default and require no configuration. Remove comparison{ ... } blocks from pint config file when upgrading.
• template check is now enabled by default and require no configuration. Remove template{ ... } blocks from pint config file when upgrading.
• rate check is now enabled by default for all configured Prometheus servers. Remove rate{ ... } blocks from pint config file when upgrading.
• series check is now enabled by default for all configured Prometheus servers. Remove series{ ... } blocks from pint config file when upgrading.
• vector_matching check is now enabled by default for all configured Prometheus servers. Remove vector_matching{ ... } blocks from pint config file when upgrading.

v0.4.4

Added

• Support parseDuration function in alert templates added in Prometheus 2.32.0

v0.4.3

Fixed

• Fixed series check handling of queries with {__name__="foo"} selectors.

v0.4.2

Fixed

• Fixed template check handling of absent calls on aggregated metrics, like absent(sum(nonexistent{job="myjob"})).

v0.4.1

Added

• template check will now warn if any template is referencing a label that is not passed to absent(). Example:

  - alert: Foo
    expr: absent(foo{env="prod"})
    annotations:
      summary: "foo metric is missing for job {{ $labels.job }}"
  

  Would generate a warning since absent() can only return labels that are explicitly passed to it and the above call only passes env label. This can be fixed by updating the query to absent(foo{env="prod", job="bar"}).

v0.4.0

Added

• comparison check will now warn when alert query uses bool modifier after condition, which can cause alert to always fire. Example:

  - alert: Foo
    expr: rate(error_count[5m]) > bool 5
  

  Having bool as part of > 5 condition means that the query will return value 1 when condition is met, and 0 when it’s not. Rather than returning value of rate(error_count[5m]) only when that value is > 5. Since all results of an alerting rule expr are considered alerts such alert rule could always fire, regardless of the value returned by rate(error_count[5m]).

Fixed

• comparison check will now ignore absent(foo) alert queries without any condition.

v0.3.1

Added

• --offline flag for pint ci command.

Fixed

• Fixed template check panic when alert query had a syntax error.

v0.3.0

Added

• rule block can now specify ignore conditions that have the same syntax as match but will disable rule for matching alerting and recording rules #48.
• match and ignore blocks can now filter alerting and recording rules by name. record will be used as name for recording rules and alert for alerting rules.

v0.2.0

Added

• --offline flag for pint lint command. When passed only checks that don’t send any live queries to Prometheus server will be run.

• template check will now warn if template if referencing a label that is being stripped by aggregation. Example:

  - alert: Foo
    expr: count(up) without(instance) == 0
    annotations:
      summary: "foo is down on {{ $labels.instance }}"
  

  Would generate a warning since instance label is being stripped by without(instance).

v0.1.5

Fixed

• Fixed file descriptor leak due to missing file Close() #69.

v0.1.4

Changed

• Retry queries that error with query processing would load too many samples into memory using a smaller time range.

v0.1.3

Added

• vector_matching check for finding queries with incorrect on() or ignoring() keywords.

Fixed

• comparison check would trigger false positive for rules using unless keyword.

v0.1.2

Fixed

• # pint skip/line place between # pint skip/begin and # pint skip/end lines would reset ignore rules causing lines that should be ignored to be parsed.

v0.1.1

Changed

• value check was replaced by template, which covers the same functionality and more. See docs for details.