Elastic Apm

Elastic APM is the application performance monitoring component of the Elastic Stack, providing distributed tracing, transaction performance metrics, error tracking, and service dependency maps alongside the Elasticsearch and Kibana tools teams may already use. For teams with existing Elastic Stack investment, adding APM avoids introducing a new APM vendor — trace and metrics data lives in Elasticsearch with the same querying and visualization toolchain. Agents with access to Elastic APM can query performance data through the Elasticsearch API alongside application and log data.

Authentication follows the Elastic Stack security model — API keys or username/password for Elasticsearch and Kibana API access. The APM Server has a separate secret token for agent-to-APM-Server communication. Teams using Elastic Cloud Deployment use the cloud-managed authentication infrastructure.

Documentation is comprehensive and covers the APM agent integration, APM Server configuration, and Elasticsearch data model for APM indices. Teams already familiar with Elasticsearch will find the APM documentation builds on familiar concepts. Teams evaluating Elastic APM versus standalone APM tools (Lightstep, Datadog APM) should weigh the Elastic Stack integration value against the operational complexity of self-hosting the full stack.

The APM Server accepts trace and metric data from APM agents (language-specific SDKs). Query access uses the Elasticsearch API — APM data is stored as Elasticsearch indices, enabling agents to query trace data, service performance metrics, and error rates using standard Elasticsearch queries. The Kibana APM API provides higher-level APM-specific endpoints for service lists, transaction summaries, and service maps.

Reliability follows the Elastic Stack deployment model — self-hosted reliability is the team's operational responsibility while Elastic Cloud provides managed availability. The APM data pipeline (agent → APM Server → Elasticsearch) introduces multiple components, each of which should be monitored for availability.