Telemetry: getting started

Add an SDK, set a DSN, deploy. Your service appears in the dashboard with its first request, and alerts, root cause and load tests build on the same data.

Getting started

1. Sign in to the TaskTrooper Telemetry dashboard with your company account (SSO). The first person creates the organisation and its first project.
2. Create an ingest key. Settings → API keys. The project's DSN carries it: https://<ingest-key>@telemetry-ingest.tasktrooper.ai/<project-id>
3. Install an SDK in each service or app and pass the DSN, usually through TROOPER_DSN. See SDKs.
4. Deploy. Traces, metrics and logs travel over OTLP. The service, its endpoints and the deploy marker appear with the first request.
5. Connect a repository (optional) to get KPI and monitor suggestions from your code.

Code insights

Install the TaskTrooper GitHub App and pick repositories. TaskTrooper reads the routes, consumers and jobs in your code and matches them with production traffic.

For each critical path it proposes a KPI and the monitors behind it. You accept, edit or dismiss each suggestion; pushes to the default branch update the inventory incrementally.

KPIs and alerts

A KPI is an objective on a service, endpoint or process, for example 99.5% availability over 30 days. Alerts fire on error-budget burn rate, so a short spike and a slow leak both page at the right time.

When an incident opens, root-cause analysis walks the call tree and tells you whether it is your code or something you call. Alerts go to Slack with the cause attached; acknowledge and mute from the thread.

Load tests

Verify a staging host once (DNS TXT record or a well-known file), then define load levels and a traffic mix, or generate a scenario from real production traffic.

Runs report through the same telemetry, so the services under test show their latency and errors next to the load curve.

Next: SDKs and AI agents.