[harshrajdebug]

What this PR does:

Builds the metrics provider of an ANALYSIS stage from the resolved metrics config, so the timeout of a referenced AnalysisTemplate is used.

• In executestage/analysis.go the newMetricsProvider function now takes the *config.AnalysisMetrics value returned by getMetricsConfig and reads the provider name from it.
• In analysisprovider/metrics/factory the NewProvider function takes *config.AnalysisMetrics instead of the *config.TemplatableAnalysisMetrics type.

Why we need it:

The provider was built from the stage's own metrics entry. When that entry only references a template, its timeout is the 30s default, so Prometheus and Datadog queries ignored the timeout set in the template. Details and the reproduction are in the issue.

TestExecutor_newMetricsProvider_timeout sets a 100ms timeout once inline and once in a template, against a Prometheus server that answers after 1s, and expects context deadline exceeded in both cases. When I pass the stage entry's timeout to the provider instead, as master does, the template case waits for the 1s response and fails.

Which issue(s) this PR fixes:

Fixes #7526

Does this PR introduce a user-facing change?:

• How are users affected by this change: Metrics analyses that reference an AnalysisTemplate now use the timeout set in the template. A timeout written next to template in the stage is no longer used, like the other fields there.
• Is this breaking change: No
• How to migrate (if breaking change): N/A

Screenshots/Videos (for documentation or website changes):

N/A

[netlify]

✅ Deploy Preview for pipecd-site canceled.

Name	Link
🔨 Latest commit	84bdde4
🔍 Latest deploy log	https://app.netlify.com/projects/pipecd-site/deploys/6adb4d0d88ce24bd272ed63a