Collect CI performance metrics on per-target level #2044
Comments
fweikert
added a commit
to fweikert/continuous-integration
that referenced
this issue
Sep 30, 2024
Previously we hard-coded non-test steps such as Buildifier, which can cause problem when adding new steps. With this commit we detect test steps by looking at their command, which has to contain `bazelci.py runner`. Progress towards bazelbuild#2044
fweikert
added a commit
that referenced
this issue
Oct 1, 2024
Previously we hard-coded the names of non-test steps such as Buildifier, which can cause problem when adding new steps. With this commit we detect test steps by looking at their command, which has to contain `bazelci.py runner`. Progress towards #2044
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Thanks to sharding it's really hard to find the performance of a single test, especially across multiple builds. Moreover, thanks to bazel-diff the target might not have been run at all.
We're already analyzing test_bep.json when --print_shard_summary is set. Consequently, it should be easy to dump per-target metrics into some database (Cloud SQL or whatever).
The text was updated successfully, but these errors were encountered: