What it is

Pinpoint is an open-source Application Performance Management (APM) tool built for large-scale distributed systems. It is written primarily in Java and also supports applications written in PHP and Python through a separate agent repository. Inspired by Google's Dapper, it provides a way to analyze the overall structure of a system and how the components inside it are interconnected by tracing transactions across distributed applications. The project lives in the Java ecosystem, is licensed under Apache-2.0, and has been developed for around twelve years, with its current stable release being v3.1.0.

The concrete problem Pinpoint addresses is visibility. Modern services are made up of many components that communicate with each other and call external services, and how any individual transaction actually executes is often left as a black box. Pinpoint traces transaction flows between those components so that problem areas and potential bottlenecks become identifiable rather than guessed at. It gives teams a topology view of the system, real-time monitoring, and code-level visibility into every transaction, and it does so by installing APM agents without changing a single line of application code, adding roughly a 3% increase in resource usage.

Key capabilities

• ServerMap visualizes how the components of a distributed system are interconnected, and clicking a node reveals details such as its current status and transaction count.
• Realtime Active Thread Chart monitors active threads inside applications in real time.
• Request/Response Scatter Chart shows request count and response patterns over time, and transactions can be selected for more detail by dragging over the chart.
• CallStack provides code-level visibility into every transaction in a distributed environment, identifying bottlenecks and points of failure in one view.
• Inspector exposes application details including CPU usage, memory and garbage collection, TPS, and JVM arguments.
• URI-metric, Infrastructure, and Error-Analysis views extend the same monitoring surface to per-URI metrics, infrastructure metrics, and error investigation.

Who uses it and how

• Teams operating large-scale distributed Java systems use the agent to trace transactions across services without modifying application source code.
• Organizations running PHP or Python applications alongside Java services use the separate c-agent repository to cover those runtimes as well.
• Platform engineers deploy Pinpoint to Kubernetes using the dedicated pinpoint-kubernetes project.
• Operators monitor active threads and request/response patterns