←Back to ProjectsPulse cover image

2026

Active

Pulse

Stack

Local Java APM agent for Spring Boot 3+ with SQL, HTTP, and JVM metrics exposed through a built-in offline web UI.

#java #apm #observability #monitoring
Repository Live Demo

Project Context

Pulse was built around a simple constraint: keep application diagnostics local.

Many monitoring products assume an external collector, a hosted dashboard, or another service to operate. Pulse takes the opposite direction. It attaches directly to the JVM as a -javaagent, starts with the application, collects useful telemetry, and exposes everything through a local web UI.

The result is a lightweight APM workflow designed for developers and teams who want visibility without adding a full observability platform.

What Pulse Collects

SQL activity

  • JDBC query instrumentation for Statement and PreparedStatement
  • query duration, errors, and SQL operation type
  • Spring context enrichment
  • SQL to HTTP correlation through traceId or thread/time fallback

HTTP activity

  • endpoint transactions with method and path
  • status code, latency, and error tracking
  • sampled hotspots and call-stack visibility
  • request metadata with masking and sensitive filtering

JVM metrics

  • heap usage
  • process CPU
  • threads and garbage collection activity
  • average HTTP latency and p95 indicators

Operating Model

Pulse is designed to stay simple in production-like environments:

  • no external collector
  • no additional database
  • no mandatory cloud dependency
  • a single agent JAR started with the target JVM

That makes it useful for local debugging, offline analysis, staging validation, and self-hosted environments where telemetry should not leave the machine.

Launch Model

The integration model is intentionally small:

java \
  -javaagent:/path/to/pulse/target/pulse-1.0.0-agent.jar=port=17321 \
  -jar /path/to/your-app.jar

Once the application starts, Pulse exposes a local UI and API on the configured port.

Why This Project Matters

Pulse is not trying to replace a full enterprise observability stack. Its value is elsewhere:

  • immediate visibility with minimal setup
  • easier diagnostics for Spring Boot applications
  • better control over sensitive HTTP and SQL metadata
  • offline-first operation for restricted environments

This makes it a strong fit for teams that want a practical troubleshooting layer before introducing heavier infrastructure.

Current Direction

The project is focused on keeping the runtime small while improving the quality of collected insights:

  • better correlation between HTTP requests and SQL execution
  • clearer hotspot and call-stack views
  • stable JVM-level metrics for day-to-day diagnostics
  • a local UI simple enough to use without onboarding overhead