search

aquanauts / alogamous

A log anomaly detection framework
MIT License
1 stars 9 forks source link

Flag stack traces that don't result in the service dying #6

Closed laurenassour closed 2 months ago

laurenassour commented 3 months ago

In our logs this is easy because we have a 'starting' header that makes it obvious when a service starts up, at least. We want to see if we're dumping stack traces that are going unnoticed.

kitrady commented 3 months ago

What is a stack trace? What is dumping a stack trace and why would it going unnoticed be a problem?

benrady commented 3 months ago

Great question! "Stack trace" is a generic name...Python calls these "Tracebacks" for some reason.

https://stackoverflow.com/questions/1508467/log-exception-with-traceback-in-python

benrady commented 3 months ago

Basically, it's something you can write to the log file that says an error occurred, what the error was, and what lines of code were involved in the error. It prints out the entire call stack (function A, calls function B, calls function C, etc...). Hence the name "stack trace"

kitrady commented 2 months ago

I will work on this