Theoretically it should be possible to correlate deployed code changes with the logs and traces preceding 500 errors; and then recreate the failure condition given a sufficient clone of production (in CI) to isolate and verify the fix before deploying new code.
Practically then, each PR generates logs, traces, and metrics when tested in a test deployment and then in production. FWIU that's the "personal" part of sprkl.
westurner|10 months ago
/? vscode distributed tracing: https://www.google.com/search?q=vscode+distributed+tracing :
- jaegertracing/jaeger-vscode: https://github.com/jaegertracing/jaeger-vscode
/? line-based display of distributed tracing information in vs code: https://www.google.com/search?q=line-based%20display%20of%20... :
- sprkl personal observability platform: https://github.com/sprkl-dev/use-sprkl
Theoretically it should be possible to correlate deployed code changes with the logs and traces preceding 500 errors; and then recreate the failure condition given a sufficient clone of production (in CI) to isolate and verify the fix before deploying new code.
Practically then, each PR generates logs, traces, and metrics when tested in a test deployment and then in production. FWIU that's the "personal" part of sprkl.
arthurgousset|10 months ago