Skip to content

Harden API consumption log collection against continuation timeouts #41676

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
pelikhan merged 2 commits into from
Jun 26, 2026
Merged

Harden API consumption log collection against continuation timeouts #41676

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
f8dd404
Initial plan
Copilot Jun 26, 2026
9996586
Harden API consumption log continuation handling
Copilot Jun 26, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion .github/workflows/api-consumption-report.lock.yml
View file
Open in desktop

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.

8 changes: 5 additions & 3 deletions .github/workflows/api-consumption-report.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -68,9 +68,11 @@ Use the `agentic-workflows` MCP `logs` tool:
- **Incremental** (history is already rich per the threshold in **Step T1**): `logs(start_date="-1d")`
- **Backfill** (first run, cache miss, or sparse history per the threshold in **Step T1**): `logs(start_date="-90d")`
- The `continuation` field is authoritative. If it is missing or `null`, stop paging even if the returned run count exactly matches your requested count.
- If `continuation` is present, make at most **2** additional continuation calls using the returned parameters. Do **not** invent your own `before_run_id` from the earliest run in the batch.
- If any continuation call times out, returns `ECONNREFUSED`, or otherwise fails once, stop collecting and proceed with the partial dataset you already have.
- Never use Bash/CLI pagination loops, sleep-based retries, or ad-hoc count/timeout tuning to chase more pages.
- Treat the run directories already present under `/tmp/gh-aw/aw-mcp/logs/` as the authoritative collected dataset. Successful continuation calls should add to that dataset; if a later continuation attempt fails, keep using the directories that are already on disk.
- If `continuation` is present, make at most **2** additional continuation calls using the returned parameters. After each successful continuation call, re-read that response's `continuation` field before deciding whether to continue. Do **not** invent your own `before_run_id` from the earliest run in the batch.
- If a continuation call times out, returns `ECONNREFUSED`, or otherwise fails with a transient tool/transport error, retry that **same** continuation call up to **2** more times with short backoff (about 15s, then about 45s). Do **not** change the returned pagination parameters, timeout, or count while retrying.
Comment on lines +72 to +73
- If the continuation call still fails after those bounded retries, stop collecting, proceed with the logs already downloaded to `/tmp/gh-aw/aw-mcp/logs/`, and clearly note in the final discussion that continuation retries were exhausted and the dataset may be partial.
Comment on lines +71 to +74
- Never use Bash/CLI pagination loops, unbounded retries, or ad-hoc count/timeout tuning to chase more pages.

Record which mode you used (`incremental` vs `backfill`) and the chosen `start_date` in Step 6 (the discussion "Cache Memory Status" details block).

Expand Down