Documentation

You’re in the Ollie reference docs. The site root has the project pitch; this section is the user-facing reference for the v0.3 dev preview.

Start here

Brand new? → Getting started walks from a fresh Kind cluster through deploying nginx and seeing per-pod HTTP metrics on the agent’s :9090 scrape endpoint, all with K8s identity attached.

Want to understand what’s actually shipping? → What works today is the honest v0.3 inventory — what’s captured, what the metric names look like, what isn’t built yet.

Curious about the architecture? → Architecture covers the sibling-container model: the upstream OBI image runs as a sibling to the agent, the agent writes OBI’s config + receives its OTLP + re-exposes everything on a single Prometheus URL.

Looking ahead? → Roadmap lays out v0.4 (control plane), v0.5 (in-cluster store + HPA), and v0.6 (TLS + cardinality controls).

Reference index

Getting things done

• Getting started — deploy on Kind, see L4 + L7 metrics from a real workload in under 10 minutes.
• What works today — feature snapshot for v0.3 with the metric names you’ll actually see.

Concepts and design

• Architecture — sibling-container model, OBI’s role, the agent’s role, the OTLP → Prometheus path.
• Roadmap — what’s coming v0.4–v1.0.

Working on it

• Contributing — branch model, commit style, where to file issues.

Deeper reading

The repo’s docs/design/ directory holds the full architectural design log: per-subsystem design docs, the architectural decision record (decisions.md — 21 ADRs and counting), and the agreed requirements (docs/requirements.md). Read those if you’re considering contributing or want to understand why the current architecture is what it is.