Index of every page on this site: llms.txt
All skills
posthog avatar

/linking-errors-to-traces

@5fe8d7d official
by posthogposthog/posthog40k stars
3,512
Run it yourself

Links a PostHog error occurrence or session's errors to OpenTelemetry traces, logs with the same trace ID, and Session Replay. Use when the user asks for trace or span IDs on an error, same-trace logs, or backend traces for errors in a session. Starts from an occurrence UUID, session ID, or error tracking issue URL. For a general issue diagnosis or replay summary, use investigating-error-issue or investigating-replay.

  • 3 files
  • 6.7 KB
  • Updated 2 days ago
  • GitHub
Use this Skill: https://skilld.dev/gh/posthog/posthog/linking-errors-to-traces

Nothing lands on disk. Nothing to clean up.

Fork this Skill

Edit a local copy. It keeps the author and licence.

SKILL.md

≈112 tokens for metadata: the name and description. ≈775 when used: this file. ≈820 more on demand in 2 files.

Description uses 5.6% of example budget

Before choosing a Skill, your Agent reads its name and description. All available Skills share that space.

  • A shorter description leaves more room for other Skills. This entry exceeds our 1% size suggestion.

In our Claude Code example, all Skill names and descriptions share 8,000 characters. This Skill uses ≈448 characters, or 5.6%.

The 1% threshold is a size suggestion. Longer descriptions can still fit.

Your model, settings, and other Skills decide how much text your Agent can read.

Example settings and source

The example uses a 200k-token context and default Claude Code settings. The count includes the name, description, separators, and when_to_use when present. Codex also counts local file paths.

Skit's source and limits: Codex 0.160.1, Claude Code 2.1.292.

Link errors to traces and logs

Use a shared ID to connect records. A matching service and time do not prove that a log or span belongs to an error.

Find the exception

For an issue ID or URL, call posthog:query-error-tracking-issue-events with include: ["exception", "correlation"], filterTestAccounts: false, and a small limit. Choose one occurrence unless the user asks for several. Say when the result is a sample.

For an occurrence ID or session ID, read finding-exceptions.md for the exact lookup. Use the user's time range when given. If no event appears, state the window you checked. Do not claim that an event never existed based on a bounded search.

Follow the IDs

  1. Replay: If the event has $session_id, call posthog:session-recording-get with that ID. found: false means no recording was found; the error or session may still exist. Use posthog:generate-app-url with /replay/{id} when a recording exists.
  2. Exact trace: For each unique, nonzero, 32-character hex $trace_id, call posthog:apm-trace-get with that ID and excludeAttributes: true. Use its _posthogUrl for the trace link. If the event has $span_id, compare it with returned span IDs as hex, ignoring case. An empty result means no retained span was found for that trace ID. Logs may still exist.
  3. Session trace candidates: If an event lacks $trace_id and the user asked about session traces, use posthog:apm-attributes-list to find a span attribute that stores the session ID. If a clear key exists, call posthog:query-apm-spans with an exact span_attribute filter, a short window around the error, flatSpans: true, and a small limit. This finds matching child spans as well as root spans. Call posthog:apm-trace-get for a candidate that needs a link. Label these as session matches. Time alone does not prove which request raised the error.
  4. Same-trace logs: When an event has a trace ID, read finding-trace-logs.md for the bounded service lookup and exact log filter. Include every severity, including info. A trace can have logs even when no span was retained.

Use posthog:generate-app-url with /error-tracking/{id} when the issue ID is known. Do not build PostHog links by hand.

Report the result

For each checked occurrence, show the error time, IDs used for the join, and links to the issue, replay, and retained trace when available. Summarize matching logs with their time and severity. Exclude secrets and unneeded personal data from log excerpts.

Name each gap separately: no trace ID on the event, trace ID but no retained span, no session attribute on spans, no recording, or no logs with that trace ID in the checked window. A session-matched span is a candidate until an exact trace or span ID connects it to the error. State any row limit or time range that kept the search from being complete. If a lookup fails, say that its result is unknown instead of reporting that no matching data exists.

Source: SKILL.md on GitHub

Skill behaviors

No rule matched.

skilld matched fixed text patterns in SKILL.md and file names. Patterns miss obfuscated code.

skilld run checks every file with the same patterns. It asks for approval before it loads a Skill with a behavior marked Needs approval.

Third-party checks

No third-party reports yet.

Provenance

Signed by skilld at 5fe8d7d. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub yesterday.

Activeupdated 2 days ago

README badge

README badge for posthog/posthog/linking-errors-to-traces