# Get started with auto-instrumentation

Auto-instrumentation injects an [OpenTelemetry](https://opentelemetry.io/) agent into your application at startup.
The agent hooks into popular libraries and frameworks to collect [traces](/observability/tracing.md) and runtime metrics — without code changes. Log export via OpenTelemetry is [available as an opt-in](/observability/reference/auto-config.md#logs-auto-instrumentation).

Supported runtimes: **Java/Kotlin**, **Node.js**, **Python**, **.NET**, and **SDK-only** mode for Go and other languages that provide their own OpenTelemetry setup.

## What you get

Once enabled, the agent automatically instruments:

| Category | Examples (vary by runtime) |
| --- | --- |
| HTTP servers | Spring MVC, Ktor, Express, Flask, Koa |
| HTTP clients | OkHttp, Apache HttpClient, `node-fetch`, `requests` |
| Databases | PostgreSQL (JDBC, node-postgres), Valkey/Redis |
| Messaging | Kafka producers/consumers |
| gRPC | Client and server calls |
| Frameworks | Spring Boot, Ktor, Micrometer bridge |

This data powers the [Nais APM](https://grafana.dev-nais.cloud.nais.io/a/nais-apm-app) dashboards — service inventory, RED metrics (Rate/Errors/Duration), dependency maps, and runtime metrics — with no extra configuration.

## Enable auto-instrumentation

Add the following to your `nais.yaml` and deploy:

**Java / Kotlin**

```yaml
spec:
  observability:
    autoInstrumentation:
      enabled: true
      runtime: java
```

**Node.js**

```yaml
spec:
  observability:
    autoInstrumentation:
      enabled: true
      runtime: nodejs
```

**Python**

```yaml
spec:
  observability:
    autoInstrumentation:
      enabled: true
      runtime: python
```

**.NET**

```yaml
spec:
  observability:
    autoInstrumentation:
      enabled: true
      runtime: dotnet
```

**SDK only (Go, etc.)**

This sets up the OpenTelemetry environment variables but does **not** inject an agent.
Use this when your application configures the OpenTelemetry SDK itself.

```yaml
spec:
  observability:
    autoInstrumentation:
      enabled: true
      runtime: sdk
```

## Verify it works

After deploying, check that traces are flowing:

1. Open the [Nais APM service inventory](https://grafana.dev-nais.cloud.nais.io/a/nais-apm-app)
2. Find your application in the list — it should appear within a few minutes
3. Click through to see RED dashboards, operations breakdown, and dependencies

If your app does not appear, check the [auto-instrumentation configuration reference](/observability/reference/auto-config.md) for troubleshooting environment variables, and make sure your application is receiving traffic.

You can also verify in [Grafana Explore](https://grafana.dev-nais.cloud.nais.io/explore) using the Tempo data source:

```
{resource.service.name="<your-app-name>"}
```

## Next steps

- [Add custom spans and metrics](/observability/tutorials/opentelemetry-instrumentation.md) for business-specific tracing
- [Explore traces in Grafana Tempo](/observability/tracing/how-to/tempo.md)
- [Avoid false errors from expected 4xx responses](/observability/tracing/how-to/avoid-false-errors.md) — fix inflated error rates in APM
- [📚 Auto-instrumentation configuration reference](/observability/reference/auto-config.md)