APM
Measure your applications' performance with OpenTelemetry. Services, latency, grouped errors, end-to-end traces, dependencies, metrics, and logs.
On this page
The APM (application performance monitoring) shows how your applications behave on the inside: how many requests they receive, how long they take, which ones fail, and whom they call. It uses OpenTelemetry, the open instrumentation standard, so it works for any language, cloud, your own server, IBM i, or browser.
Your applications send their telemetry to ARGUNIS with a token from your company. From that moment, services appear automatically in the APM menu.
What you get
What | Where you see it |
|---|---|
Services discovered automatically when their first trace arrives, grouped by system, with their health over the last few minutes | APM → Services |
Performance of each service: requests, error rate, and p50, p95, and p99 latency, compared with 7 days ago | Service page |
Endpoints with their traffic, errors, and latency, and the traces for each one | Service page, Endpoints and traces tab |
Grouped errors: each exception only once, with how many times it occurred, since when, in which version, and a sample trace | APM → Errors |
End-to-end traces: from the click in the browser to the back end and the database, with the critical path | APM → Traces |
Dependencies: which databases, APIs, queues, and services each one calls, and who calls it | APM → Dependency map, APM → Live APM (also on the NOC screen), and the service page |
Application metrics sent by your code or its runtime (JVM, connection pool, memory…) | Service page, Charts tab |
Warning and error logs, linked to their trace | Service page, Errors and logs tab, and each trace |
Alerts for high latency, high error rate, a service that stops sending data, and new errors, which open incidents in Core | APM → Alerts |
The APM menu appears if your company's plan includes the module. The plan can also set how much telemetry per month and how many services it includes: usage is shown in APM → Configuration and settings.
Pages in this section
What you can monitor with APM: languages, versions, eBPF, and legacy systems.
Instrument a server application: Node.js, Python, Java, .NET, PHP, Go, and Ruby.
Instrument in clouds and containers: AWS, Azure, Google Cloud, Railway, Kubernetes, Docker, and the Collector.
If this is your first time, follow the order: enable the APM, create a token, instrument an application, and wait for it to appear in APM → Services.