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