Documentation Index
Fetch the complete documentation index at: https://docs.livepeer.org/llms.txt
Use this file to discover all available pages before exploring further.
Why a guide on observability
The protocol stores its rules and balances on Arbitrum One. The network’s actual work happens off-chain across orchestrators, gateways, and pipelines that the chain never sees. Reading the network well requires combining on-chain state with off-chain signal - and knowing which signal each surface gives you. A founder evaluating Livepeer needs different data than an orchestrator monitoring its own performance, or a developer integrating a metrics widget into their site. This guide maps the available surfaces to the questions they answer.The five public surfaces
Each surface has a strict scope. The Explorer cannot tell you whether an orchestrator is currently online. The Grafana dashboards cannot tell you who voted on a LIP. The capabilities API cannot tell you historical revenue. Use the surface that matches the question.Livepeer Explorer
The Explorer is the canonical view of on-chain protocol state. Everything the protocol contracts hold is queryable here.Subgraph and historical analytics
The Livepeer subgraph indexes every on-chain event into a queryable schema. It is the right surface for time-series questions, custom dashboards, and any analysis that needs more than current state. Common analytics patterns:- Network usage - segments transcoded, AI inference calls, total minutes per period.
- Fee revenue - ETH redeemed by orchestrators per period, demand-side fees per gateway.
- Orchestrator performance - per-orchestrator round-by-round reward calls, fee earnings, stake changes.
- Governance participation - votes cast per LIP, quorum trajectories, orchestrator vs delegator participation rates.
- Treasury activity - treasury inflows from inflation, outflows from passed proposals, balance trajectories.
Network dashboards and the leaderboard
Off-chain network performance is published through Grafana dashboards and a performance leaderboard. These surfaces answer questions the chain cannot - latency, throughput, success rate per pipeline, geographical distribution of work. These surfaces have a structural blind spot: they only see what their telemetry covers. An orchestrator that does not publish telemetry is invisible to the dashboards even when it is serving production traffic. Combine off-chain dashboards with on-chain data to see the full picture.Capabilities API
Each orchestrator publishes its current capabilities as a real-time advertisement: video transcoding profiles, AI inference pipelines, BYOC containers, real-time AI sessions, and the prices for each. Gateways read this advertisement to assemble a working set per job. For an external observer or integrator, the capabilities API is the surface to use to answer:- Which orchestrators currently advertise a given AI pipeline.
- What prices orchestrators are quoting for transcoding profiles.
- Where the network has spare capacity for a specific workload.
- Whether a BYOC container is being run by anyone in the active set.
Embedding live network data
Several of the surfaces above are designed to be embedded in external sites - a portfolio tracker, an orchestrator’s transparency page, a public ecosystem dashboard.Where to go next
Every Explorer metric explained, with formulas for institutional analysis.
Read paths, surface trade-offs, and off-chain blind spots in detail.
Live on-chain state, governance proposals, and per-orchestrator data.
Indexed historical events for custom analytics and embedded dashboards.