Why is my logging flooding production logs after deploy and how to fix it?

[AbigailLW]

I just ran into a weird issue where my app's logging system started outputting thousands of lines of debug data to the production console, completely swamping the actual error messages. I'm pretty sure it's a problem with the configuration file not being parsed correctly after the last deployment, but I can't pinpoint why the default fallback settings aren't kicking in. Has anyone else dealt with a logging framework going haywire like this?

[AveryQA]

I ran into something similar after a deployment last year. Sudden flood of debug lines drowned the real errors. Turned out the production config had a different root level than expected, so the app started at debug everywhere. We bumped the root level to WARN in prod and rolled that change out, and the noise dropped.

[Chloe57]

I remember chasing a path issue too. There was a symlink in the deploy that pointed to a stale config, so the defaults were never applied and the log output stayed loud. We ended up restarting with the real file in place and it calmed down, but you could still see a few stray messages.

[Gary32]

One quick move we tried was adding a light-weight sampling rule for the noisy loggers. It cut the chatter a good amount, though you still get the critical stuff. It felt temporary but it bought us time to fix the root cause.

[NatalieUW]

Did you check whether a new logger was registered at runtime that's flooding the output, instead of the config not loading?