Monitoring and Diagnosing the Performance Problems of Enterprise Applications
summary
Shows concrete diagnostics and fixes for Cloud Foundry apps: custom metrics via Graphite/InfluxDB plus CF firehose, dynaTrace and New Relic for tracing, and Liberty buildpack profiling. Root causes include thread-pool exhaustion that reduces throughput to 12 TPS with 90s latency despite low Diego cell CPU, and GoRouter behavior — GoRouter no keepalives — raising kernel CPU and connection limits. A VM change from PV to HVM increases garden container density from 30 to 200 containers and resolves a key scaling bottleneck.
AI-written summary · about
chapters (6)
- 0:00Introduction to monitoring
- 3:15Cloud Foundry monitoring tools
- 8:17Advanced Cloud Foundry metrics
- 13:43Enterprise application insights
- 16:11Diagnosing performance issues
- 24:53Summary and tools overview
related talks
52mIs your JVM app flying blind? Unmask hidden issues with Observability superpowers!
47mHow I Solved Production Issues with OpenTelemetry (And How You Can Too)
01h 28mThe dark art of performance tuning or how to become a performance hero ...
31mRunning Java Applications on Cloud Foundry
34mLive Production Debugging with Cloud Foundry and Stackdriver in 3 Easy Steps
44mReusable Insights from Enterprise Deployments
01h 00mBulletproof Java Enterprise applications for the hard production life
53mA cloud cost saving journey: Strategies to balance CPU for containerized JAVA workloads in K8s