Skip to content

For Operators

This section covers everything needed to deploy and operate the HelixObs stack for one or more instruments.

Pages in this section

Stack Architecture All services, data flows, and ports
Deployment Docker Compose setup, environment variables, volumes
Log Collection Alloy vs OTLP — what each requires and what breaks without it
Grafana Dashboards What each dashboard shows and what data it depends on
Notifications Slack and GitHub Issues — config format, rate limiting, silence rules
Authentication Setup JWT auth backends, key rotation, rollout
Adding an Instrument Instrument YAML config, auth setup, onboarding checklist

Operator responsibilities at a glance

  1. Deploy the stack — herald, OTel Collector, TimescaleDB, Loki, Tempo, Prometheus, Grafana, Alloy
  2. Configure log collection — decide per-instrument whether Alloy or OTLP delivers logs, and ensure the required pipeline is in place
  3. Configure instrument YAML — one file per instrument for notifications and auth
  4. Provide connection details to instrument teams — herald address, auth endpoint, credential
  5. Monitor — Platform Health dashboard, error rate alerts, Sherlock for AI-assisted diagnosis