Sample Traces at the Tail for Completeness
Buffer all spans and sample at the end so traces with errors are always kept and healthy traces are sampled by rate.
Task
Instrument tail_sampling: { policies: [{ name: 'errors', type: status_code, status: ERROR }] } so on-call engineers can diagnose an incident without guessing at log lines.
When to use this
An outage lasts 47 minutes because the team cannot correlate tail_sampling: { policies: [{ name: 'errors', type: status_code, status: ERROR }] } across services without structured traces.
Prerequisites
- An application with structured logging available
- A metrics or tracing backend configured
Steps
1. Configure tail_sampling: { policies: [{ name: 'errors', type: status_code, status: ERROR }] } in the project configuration file with the recommended defaults.
tail_sampling: { policies: [{ name: 'errors', type: status_code, status: ERROR }] }
Expected output: The configuration is accepted and tail_sampling: { policies: [{ name: 'errors', type: status_code, status: ERROR }] } runs without errors on the first invocation.
2. Run the check locally to verify tail_sampling: { policies: [{ name: 'errors', type: status_code, status: ERROR }] } produces the expected result before pushing.
act -j build # or: pnpm test
Expected output: Local run matches CI behavior; no surprises after push.
3. Add the gate to the CI pipeline so tail_sampling: { policies: [{ name: 'errors', type: status_code, status: ERROR }] } blocks merge on failure.
- name: Run add trace tail sampling
run: pnpm test
Expected output: Pull requests with failing tail_sampling: { policies: [{ name: 'errors', type: status_code, status: ERROR }] } are blocked from merging.
Troubleshooting
Problem
The check passes locally but fails in CI
Solution
Pin dependency versions and the runtime to match CI exactly; reproduce with the same Docker image or act.
Problem
Configuration is rejected with a schema error
Solution
Check the tool version matches the config schema; older versions may not support newer fields.
Limitations
- The gate only runs on events that trigger the workflow; locally pushed commits bypass it.
- Configuration drift is possible if someone edits settings outside version control.
Make the system observable
Start by measuring the problem. Buffer all spans and sample at the end so traces with errors are always kept and healthy traces are sampled by rate. The most common mistake is adding the gate without first confirming it catches the issue you care about. Run tail_sampling: { policies: [{ name: ‘errors’, type: status_code, status: ERROR }] } against the current codebase and inspect the output before wiring it into CI.
tail_sampling: { policies: [{ name: 'errors', type: status_code, status: ERROR }] }Correlate across services
Once the gate is in place, make failures actionable. A failing tail_sampling: { policies: [{ name: ‘errors’, type: status_code, status: ERROR }] } should point to the exact file and line, not a generic error. Pin the tool version in the lockfile so a upstream release does not silently change behavior. Review the gate output in every pull request just like a test result.