promql(histograms): scale a histogram the same as the count #16828
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
Should we go with this approach, this might actually be close to the final code, which proves the simplicity. (Assuming I haven't missed anything substantial…) Another thought: Even though this will have overblown values for some buckets, I am not so sure anymore if "fixing" those is the right thing. By reducing the overblown buckets, we are increasing other buckets, so one might argue we are actually glossing over differences between buckets that has been in the data, but has been "smoothed out" with the other approaches (3) and (4). Or in other words: The outcome of this approach (2) is easier to reason with. For "optimal" rate estimation, the hopes are that the new |
This was referenced Jul 3, 2025
This deals with the count field of native histograms in the same way as with simple float counters. It then scale the whole histogram with the same factor as it has scaled the count. This will still allow individual buckets to get extrapolated below zero, but maybe that is fine. This implements approach (2) as described in #15976 (comment) Signed-off-by: beorn7 <beorn@grafana.com>
krajorama
approved these changes
Jul 10, 2025
Konstantinov-Innokentii
added a commit
to grafana/mimir
that referenced
this pull request
Jul 17, 2025
Upstream pr - prometheus/prometheus#16828 Signed-off-by: Innokentii Konstantinov <innokenty.konstantinov@grafana.com>
Konstantinov-Innokentii
added a commit
to grafana/mimir
that referenced
this pull request
Jul 21, 2025
## Update mimir-prometheus dependency *This PR was automatically created by the [update-vendored-mimir-prometheus.yml](https://github.com/grafana/mimir/blob/main/.github/workflows/update-vendored-mimir-prometheus.yml) workflow.* ### Details: - **Source branch/ref**: [`main`](https://github.com/grafana/mimir-prometheus/tree/main) - **Previous commit**: [`73cbc98c25ff`](grafana/mimir-prometheus@73cbc98c25ff) - **New commit**: [`acace3cd1e0232d0d167e1f8062accd0a318d06c`](grafana/mimir-prometheus@acace3c) - **Changes**: [`73cbc98c25ff...acace3cd1e0232d0d167e1f8062accd0a318d06c`](grafana/mimir-prometheus@73cbc98...acace3c) ### Note to reviewer: There are two manual changes to make it mergeable: - [Use NewLeveledCompactorWithOptions instead of NewLeveledCompactorWithChunkSize](7e0061b) – NewLeveledCompactorWithChunkSize was removed in upstream - [ Port "Prevent extrapolation below zero for histogram count" to MQE ](https://github.com/grafana/mimir/pull/12106/files#diff-1256527183ea6d432ef0f080e75927eefb108f47554f9e20633c1b30756a3134). This one is more subtle. prometheus/prometheus#16828 was merged in upstream, affecting query results and causing some MQE tests to fail. The solution was to port the change to the MQE histogram rate implementation. --------- Signed-off-by: Innokentii Konstantinov <innokenty.konstantinov@grafana.com> Co-authored-by: mimir-github-bot[bot] <mimir-github-bot[bot]@users.noreply.github.com> Co-authored-by: Innokentii Konstantinov <innokenty.konstantinov@grafana.com> Co-authored-by: Dimitar Dimitrov <dimitar.dimitrov@grafana.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This deals with the count field of native histograms in the same way
as with simple float counters. It then scale the whole histogram with
the same factor as it has scaled the count. This will still allow
individual buckets to get extrapolated below zero, but maybe that is
fine.
This implements approach (2) as described in
#15976 (comment)
This is now ready for review.
For posterity: This approach was chosen as per discussion in #15976 , with @gen1321 and @krajorama being the other two contributors.