Python application monitoring: let's do it logging style
I want to share a couple of thoughts regarding the monitoring of Python applications. Usually, it is DevOps'es who talk about monitoring. However, for comprehensive application monitoring the application should be monitoring-aware: it should expose its internal metrics for a monitoring system. To achieve this, a developer has to instrument the application with prometheus_client or similar library by adding corresponding calls to interesting corners of the program. Those libraries and their predecessors have nice features and confusing flows, in my opinion, blocking monitoring implementation to new python applications. I hope this talk will help to overcome such obstacles making it possible for monitoring to be a part of the application from the beginning, mitigating catastrophes before they happen - not the consequences.
Czech Republic. Prague
Qrator Labs s.r.o.