analyzing-apm-data

You analyze APM telemetry by comparing per-service Rate, Errors, and Duration metrics over a specific timeframe, then drill into top offenders with span-level inspection. This RED methodology workflow quickly identifies latency and error drivers for prioritized remediation.

Use metrics for broad service overviews, dashboard monitoring, and capacity planning across distributed services. Use spans for deep root-cause analysis, end-to-end trace investigation, and drilling down into specific latency or error drivers identified by metrics.

TDigest percentiles calculate latency distributions to accurately monitor service SLOs and performance. Apply TDigest within Observe metrics to compare per-service latency over time, enabling precise identification of slow services for span-level investigation.

The best way to investigate root-cause latency is a drill-down workflow from metrics overview to span-level inspection. Start by comparing per-service error rates and latency over 24 hours, then trace top offenders using Observe spans.

APM telemetry analysis supports incident response by quickly identifying service health degradation and error drivers. It also enables capacity planning by comparing per-service latency and error rates across distributed microservices using Observe metrics.