Start with aggregates and representative traces, then expand only when the evidence demands it.
The decision behind the technique
Large log dumps waste context and can expose personal data or credentials. A bounded query produces better diagnostic evidence and a smaller disclosure surface.
Start with aggregates and representative traces, then expand only when the evidence demands it.
A practical step-by-step workflow
Use this sequence as a default, then tighten or expand it only when the evidence changes.
- Set the incident time range and affected service
- Group by error signature or status code
- Select a few representative request IDs
- Redact secrets and personal data before sharing details
Failure boundary and finish line
Do not paste environment dumps, cookies, authorization headers or complete user payloads.
Finish when the tool has a narrow purpose, a trustworthy source and a clear fallback.
Frequently asked questions
What should I do first?+
Set the incident time range and affected service
What is the main failure mode?+
Do not paste environment dumps, cookies, authorization headers or complete user payloads.
OFFICIAL SOURCES
Verify against the source
OpenAI MCP guidanceSentry MCP documentationThis guide provides general information, not an OpenAI allowance promise or support decision. Your account and official pages govern actual usage, benefits and recovery times.