Skip to main content

Observability

Monitor application performance through metrics and observability data.

Overview​

The Observability API provides access to:

  • Application Metrics — Performance, throughput, latency
  • JVM Metrics — Memory, GC, threads
  • Business Metrics — Transformation counts, error rates
  • Prometheus Format — Compatible with popular monitoring stacks

Integration with Monitoring Stack​

This API integrates with:

Operations​

📄️ List available metrics

Returns the names of all Micrometer metrics currently registered with the application.

📄️ Get metric value

Returns the current value(s) for a specific Micrometer metric.

📄️ Prometheus scrape endpoint

Exposes all Micrometer metrics in Prometheus text exposition format.

Common Metrics​

transform_files_processed_total        — Total files transformed
transform_records_published_total      — Total records published to Kafka
transform_errors_total                 — Total transformation errors
transform_duration_seconds             — Duration of transformations
jvm_memory_used_bytes                  — JVM heap memory in use
jvm_threads_current                    — Active JVM threads
http_requests_total                    — Total HTTP requests
http_request_duration_seconds          — HTTP request latency

Example: Query Metrics​

# List all available metrics
curl http://localhost:8080/actuator/metrics

# Get a specific metric (JVM memory)
curl http://localhost:8080/actuator/metrics/jvm.memory.used

# Prometheus-compatible format (scrape by Prometheus)
curl http://localhost:8080/actuator/prometheus

Prometheus Configuration​

Add to prometheus.yml:

scrape_configs:
  - job_name: "transform-platform"
    static_configs:
      - targets: ["localhost:8080"]
    metrics_path: "/actuator/prometheus"
    scrape_interval: 15s

Grafana Dashboards​

Pre-built dashboards available:

  • JVM Performance
  • Application Throughput
  • Error Rates
  • Kafka Publishing Metrics

Custom Metrics​

Application publishes custom metrics:

  • transform.files.processed — Gauge
  • transform.records.published — Counter
  • transform.errors — Counter
  • transform.duration — Timer

Access via Prometheus queries:

rate(transform_files_processed_total[5m])      # 5-min throughput
transform_errors_total                          # Total errors
transform_duration_seconds_bucket{le="+Inf"}  # Duration percentiles