Build a production-ready monitoring stack for Presto using Prometheus, Grafana, Alert Manager and Slack. Learn the techniques used by engineering teams operating petabyte-scale clusters from Kiersten Stokes - Open Source Software Developer at IBM.

Workshop highlights:

Observe: Deep Dive into Presto Metrics
- JMX MBean exports and JVM internals
- Coordinator and worker health monitoring
- Resource group queue and concurrency tracking
Visualize: 24-Panel Grafana Dashboard
- Build production-grade dashboards from scratch
- Identify query bottlenecks and memory pressure
- Latency percentiles, GC overhead, and spill analysis
Alert: Production Alerting with Prometheus
- Write tiered PromQL alerting rules
- SLA breach, capacity, and failure rate detection
- Reduce noise with baseline traffic guards
Respond: Incident Response via Slack
- Trigger real alerts with live query simulations
- Color-coded Slack notifications with severity routing
- Simulate multi-tenant resource contention

Try Presto yourself, by signing up for the free watsonx.data trial or by using the open-source version here.