[llvm] [CI] Add queue size, running count metrics (PR #122714)

[Aiden Grossman via llvm-commits]

Nathan =?utf-8?q?Gauër?= <brioche at google.com>,
Nathan =?utf-8?q?Gauër?= <brioche at google.com>
Message-ID:
In-Reply-To: <llvm.org/llvm/llvm-project/pull/122714 at github.com>


================
@@ -164,16 +230,19 @@ def main():
     # Enter the main loop. Every five minutes we wake up and dump metrics for
     # the relevant jobs.
     while True:
-        current_metrics = get_metrics(github_repo, workflows_to_track)
+        current_metrics = get_per_workflow_metrics(github_repo, workflows_to_track)
+        current_metrics += get_sampled_workflow_metrics(github_repo)
         if len(current_metrics) == 0:
-            print("No metrics found to upload.", file=sys.stderr)
-            continue
+            print("No metrics found to upload.", file=sys.stdout)
 
         upload_metrics(current_metrics, grafana_metrics_userid, grafana_api_key)
-        print(f"Uploaded {len(current_metrics)} metrics", file=sys.stderr)
+        print(f"Uploaded {len(current_metrics)} metrics", file=sys.stdout)
----------------
boomanaiden154 wrote:

GCP classifies everything written to `stderr` as a sever error? I don't think I've observed that yet.

All the default logging libraries write out to `stderr` by default, including the builtin `logging` and `absl.logging`.

https://github.com/llvm/llvm-project/pull/122714