Skip to content

[FLINK-32649][metrics] Do not report a metric whose variables differ from an earlier metric of the same name - #29172

Open
MartijnVisser wants to merge 2 commits into
apache:masterfrom
MartijnVisser:FLINK-32649-label-binding
Open

MartijnVisser wants to merge 2 commits into
apache:masterfrom
MartijnVisser:FLINK-32649-label-binding

Conversation

@MartijnVisser

Copy link
Copy Markdown
Contributor

What is the purpose of the change

We store a collector under the metric name alone, so a metric whose variables differ from the first of that name reuses it and its values go out under the wrong label names. Depends on #29171.

Brief change log

  • Stop reporting a metric whose label names do not match its collector
  • Copy the variables in iteration order in the removal test, since values bind by position

Verifying this change

This change added tests and can be verified as follows:

  • metricIsNotReportedWhenItsVariablesDifferFromAnEarlierMetricOfTheSameName fails before this change and passes after it

Does this pull request potentially affect one of the following parts:

  • Dependencies (does it add or upgrade a dependency): no
  • The public API, i.e., is any changed class annotated with @Public(Evolving): yes, AbstractPrometheusReporter is @PublicEvolving. No signature changes, but a mismatching metric that used to be reported under the wrong label names no longer is
  • The serializers: no
  • The runtime per-record code paths (performance sensitive): no
  • Anything that affects deployment or recovery: JobManager (and its components), Checkpointing, Kubernetes/Yarn, ZooKeeper: no
  • The S3 file system connector: no

Documentation

  • Does this pull request introduce a new feature? no
  • If yes, how is the feature documented? not applicable

Was generative AI tooling used to co-author this PR?
  • Yes (please specify the tool below)

Generated-by: Claude Code (Claude Opus 5)

@flinkbot

flinkbot commented Sep 14, 2026

Copy link
Copy Markdown
Collaborator

CI report:

Bot commands The @flinkbot bot supports the following commands:
  • @flinkbot run azure re-run the last Azure build

… into one label name

Two metric group variables that differ only in characters the filter replaces
become the same label name. Prometheus rejects such a body and abandons the
whole scrape, so reporting the metric costs every other metric of the process.
Removal has to tolerate a metric that was never registered.

Generated-by: Claude Code (Claude Opus 5)
…from an earlier metric of the same name

A collector is stored under the scoped metric name alone, so a metric whose
variables differ from the first one registered reuses it and its values are
published under the other metric's label names.

Generated-by: Claude Code (Claude Opus 5)
@MartijnVisser
MartijnVisser force-pushed the FLINK-32649-label-binding branch from 17294dd to 0140aaf Compare September 15, 2026 11:48
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants