mirror of https://github.com/prometheus/prometheus
Browse Source
When running tests in parallel, 10 milliseconds may not be enough for all discoverers to register, which will make test flaky. This commit changes the waiting logic to wait for number of discoverers to stop increasing during given time frame, which should be large enough for single discoverer to register in test environment. A following run passes with this commit: go test -failfast -race -count 100 -v ./discovery/kubernetes/ Signed-off-by: Mateusz Gozdek <mgozdekof@gmail.com>pull/9651/head
Mateusz Gozdek
3 years ago
committed by
GitHub
1 changed files with 6 additions and 3 deletions
Loading…
Reference in new issue