Install the agent, point it at your service and read your first golden signals.

You do not need a week and a platform team to get your first useful dashboard. You need the agent, one config line, and about five minutes. Here is the whole path.

Install the agent

curl -sSL https://get.helix.dev | sh
helix agent start --service checkout

Point it at your service

The agent auto-instruments common frameworks, so within a minute of starting it you have requests, latency, and error rate flowing in with no code changes. If you already emit OpenTelemetry, it picks that up too.

Read your golden signals

Open the generated service view and you will see the four signals that answer "is it healthy?" — traffic, latency, errors, and saturation — already populated. No panel building required for the basics.

  • Latency, as a percentile distribution, not a misleading average.
  • Error rate, split by route so you can see which endpoint hurts.
  • Throughput, so the other three have context.

Then make it yours

The default view gets you oriented; from there you add the one or two panels that reflect your service's specific failure modes. But the first five minutes should be zero-effort, and now they are.