NewYour coding agent can read the release notes before it upgrades.Set up the MCP server →
Go modules
Go modules, #1810 by repository stars
Latest stable
v0.90.0
Last release 11 days ago
11 Nov 2026
Ships on a steady schedule
a new release about every 1 weeks
Nearly every release is documented
notes for 60 of the last 60 stable releases
Nothing withdrawn
no release was ever pulled
6 years old
420 releases · first in 2021
Drag across the ruler to compare, or tap a release to open it.
Added support for Prometheus v3.16.0, which adds the new experimental PromQL function integral() and makes the anchored and smoothed range selector mo
integral() and makes the anchored and smoothed range selector
modifiers stable, so they no longer require
--enable-feature=promql-extended-range-selectors.pint ci detection of git changes.Nothing published for this version
Fixed alerts/count check undercounting alerts for rules with keep_firing_for - #2004.
alerts/count check undercounting alerts for rules with
keep_firing_for - #2004.Your coding agent can read these notes before it upgrades. Set up the MCP server →
Nothing published for this version
Fixed premature cache expiry for range queries.
Nothing published for this version
Nothing published for this version
Improved Prometheus query cache efficiency.
pint_rule_file_owner metric is now exported immediately after all rules are parsed, instead of waiting for all checks to complete.
pint_rule_file_owner metric is now exported immediately after all rules are
parsed, instead of waiting for all checks to complete.sampleRateLimit option was being applied to all requests, instead of only
requests that can load samples. Now it only applies to instant and range queries.pint watch Prometheus query cache was being cleaned prematurely
while cached responses might have been still needed, resulting in unnecessary queries.
Cache evictions now happen only after each pint watch iteration.Added sampleRateLimit option to Prometheus server configuration. It limits the number of samples Prometheus reads per second to process queries sent b
sampleRateLimit option to Prometheus server configuration.
It limits the number of samples Prometheus reads per second to process
queries sent by pint. It requires query statistics
to be enabled.pint_prometheus_sample_rate_limit and pint_prometheus_query_samples_total
metrics for the sampleRateLimit option.promql/rate check now reports increase() calls on non-counter metrics
and validates the time window passed to increase(), same as for rate()
and irate() - #1989.Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Added support for Prometheus v3.14.0 experimental PromQL function: start_timestamp().
start_timestamp().first_over_time() is now stable as of Prometheus v3.14.0 and no longer
requires --enable-feature=promql-experimental-functions.--enable-feature=promql-duration-expr.Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Added support for Prometheus v3.13.0 experimental PromQL functions: min_of(), max_of().
min_of(), max_of().query template function, reporting syntax errors and result
labels that the query cannot produce.tmpl template function.Nothing published for this version
Nothing published for this version
Nothing published for this version
promql/vector_matching was refactored heavily to be faster and more accurate.
abs(), clamp_min(), hour(), ln(),
histogram_quantile(), or timestamp() wrap aggregations that already remove
labels.name regexp is not set.$annotations map, causing template expansion to always return
empty values for annotation references.# pint disable promql/series comments having no effect
on promql/series checks.pint_rule_file_owner metric using symlink target path instead
of symlink name, making it inconsistent with pint_problem metric.Added promql/offset check that warns when a PromQL query uses an offset larger than configured Prometheus retention.
offset larger than configured Prometheus retention.start(), end(), range(), step().sum by(job) vs sum without(instance)).Nothing published for this version
GitHub and GitLab reporters will now automatically delete outdated general comments from previous pint ci runs.
pint ci runs.Nothing published for this version
rule/dependency check now supports ignoreGroupMismatch option for ignoring specific metrics when checking for cross-group dependencies - #1795.
ignoreGroupMismatch
option for ignoring specific metrics when checking for cross-group
dependencies - #1795.Fixed incorrect reports from promql/regexp when empty match group is used.
rule/dependency won't report warnings when one query uses a recording rules that produce a static value (vector(1)). If the value is always the same t
vector(1)).
If the value is always the same the lag doesn't matter and it's a false positive.vector(1)).
If the value is always the same the lag doesn't matter and it's a false positive.GitHub and GitLab reporter was trying to create comments on unmodified files, which is not allowed.
rule/dependency cross-group dependency check now only reports problems for recording rules.
pint ci.Fixed GitLab reporter sending invalid line position when a comment is placed on a modified line in the diff, causing a 400 error from GitLab API.
When running pint ci pint will now use git diff to calculate modified lines instead of git blame.
pint ci pint will now use git diff to calculate modified lines
instead of git blame.\r\n line endings correctly.rule/dependency check will now warn when a rule uses a metric produced by a recording rule from a different group, which can introduce evaluation lag.
scalar(), vector(), time(), pi(), or number literals.Nothing published for this version
Nothing published for this version
Nothing published for this version
Removed uber-go/automaxprocs as it's no longer needed with Go 1.25+.
Nothing published for this version
Nothing published for this version
Nothing published for this version
Fixed false positive reports from promql/impossible when label_join or label_replace is used - #1631.
label_join or label_replace is used - #1631.Nothing published for this version
Nothing published for this version
Fixed a panic when parsing incomplete rules using relaxed mode.
Nothing published for this version
The promql/impossible check will now warn about aggregations and and binary operations that try to operate on labels that are already removed in the q
The promql/impossible check will now warn about aggregations and and binary operations that try to operate on labels that are already removed in the query. Example:
- record: foo
expr: sum(foo) / on(cluster) sum(bar)
The above tries to join two series on the cluster label, but sum(...) already removed labels
from the results on both side, so there will be no such label on join on.
The promql/impossible check will now warn about label joins
using group_left(...) and group_right(...) that are not possible or not used in any way.
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
CI comments would sometimes print the problematic query fragment twice.
Nothing published for this version
Added support for toDuration and now template functions added in Prometheus v3.6.
toDuration and now template functions added in Prometheus v3.6.One column per quarter.