It has been a busy several years in monitoring and observability. Logging solutions are widely deployed, from scp+grep to proper federated logging. Collecting centralised server environment and resources metrics has also matured significantly. In this talk we'll introduce the next layer in this onion - in-depth Application Performance Monitoring (APM). While other parts of the monitoring stack could be described as watchpeople, APM plays much more of a detective role, bringing ever more metrics into increasingly complex UIs to suss out performance bottlenecks.
We'll touch on the history of APM and show a demo of Elastic's APM, hunting down a bug that's tricky to diagnose with logs and metrics alone. We'll also discuss how web tech-driven organisations can get more out of their APM systems with 2 real case studies. We'll finish by discussing vendor-agnostic principles for evaluating your troubleshooting needs and ensuring the tools you use continue to fit your team.