Azure / ARO-RP / 27609736022

39%

Build Type

push

github

Committed by web-flow

Commit Message

[Critical] fix(monitor): tighten timeout cleanup and close idle connections (#4882)

* fix(monitor): close cluster monitor idle connections

Make cluster monitors retain the shared Kubernetes HTTP client so the monitor
can close idle connections after each cycle instead of leaking transport state.

Co-authored-by: Cursor <cursoragent@cursor.com>

* fix(monitor): bound worker cleanup after timeouts

Let timed out monitor cycles wait briefly for in-flight goroutines and then
close closeable monitors so transport-backed workers do not linger indefinitely.

Co-authored-by: Cursor <cursoragent@cursor.com>

* fix(monitor): close aro client idle connections

Close the ARO typed client's idle HTTP connections when cluster monitors are disposed so the goroutine cleanup change also covers the separate arocli transport. Add targeted tests for the real close path and the fake clientset nil REST client guard.

Co-authored-by: Cursor <cursoragent@cursor.com>

* test(monitor): align cleanup test with package style

Use the standard library panic check in the fake ARO client cleanup test and document the typed-nil REST client guard so the cleanup path stays consistent with nearby monitor tests.

Co-authored-by: Cursor <cursoragent@cursor.com>

* docs(monitor): clarify closeable cleanup contract

Document that Closeable implementations must tolerate forced cleanup before
Monitor returns, including concurrent and repeated Close calls.

Co-authored-by: Cursor <cursoragent@cursor.com>

* fix(monitor): distinguish timeout from cancellation in workOne

Differentiate context.DeadlineExceeded from parent context cancellation
so the timeout metric and log message are only emitted on actual
timeouts, not on graceful shutdown.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

---------

Co-authored-by: Cursor <cursoragent@cursor.com>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Coverage Stats

36 of 46 new or added lines in 2 files covered. (78.26%)

5 existing lines in 1 file now uncovered.

30376 of 77868 relevant lines covered (39.01%)

0.44 hits per line