Grafana Dashboards for Spring Boot Service Health

A Grafana dashboard for a Spring Boot service should answer four questions at a glance — the four golden signals: latency, traffic (request rate), errors, and saturation. Every other panel is a supplement to these four.

Four golden signals

• Request rate: rate(http_server_requests_seconds_count{application="my-service"}[5m]) — requests per second; visualize as a time-series line; spike indicates load increase
• Error rate: rate(http_server_requests_seconds_count{status=~"5.."}[5m]) / rate(http_server_requests_seconds_count[5m]) — percentage of 5xx responses; alert when above 1%
• Latency: histogram_quantile(0.99, rate(http_server_requests_seconds_bucket[5m])) — P99 latency in seconds; alert when above SLO threshold (e.g., 500ms)
• Saturation: JVM heap usage jvm_memory_used_bytes{area="heap"} / jvm_memory_max_bytes{area="heap"} and HikariCP pending connections hikaricp_connections_pending

JVM health panels

• Heap usage over time: identify trends toward GC pressure before the GC becomes a problem; a sawtooth pattern is normal GC behavior; a flat rising line without drops indicates a memory leak
• GC pause duration: rate(jvm_gc_pause_seconds_sum[5m]) / rate(jvm_gc_pause_seconds_count[5m]) — average GC pause time; above 200ms on average indicates GC tuning is needed
• Thread count: jvm_threads_live_threads trending upward indicates a thread leak

Database health panels

• HikariCP active connections: hikaricp_connections_active — approaching maximumPoolSize indicates pool saturation
• Connection acquisition time: how long threads wait for a pool connection; above a few milliseconds consistently means queries are too slow or pool size is too small

Organize the dashboard into rows: Service Health (golden signals), JVM Health, Database, and Infrastructure. Use Grafana's variable feature for $application and $environment to make one dashboard serve all services.