Usage and cost

Home is the Rius workspace’s summary page: how your agents did this period, in volume, cost, tokens, and errors.

The Home page for a workspace: trace count against the previous period, a needs-you-now list, KPI tiles for spans, LLM calls, duration, cost and tokens, then charts for volume, errors, latency, time to first token, cost by model, and top exceptions

The stat tiles compare the selected window (15m / 1h / 1d / 7d / 30d / custom) against the previous one: total traces, average duration, total cost with per-trace and per-1k-token averages, and total tokens. Below them, panels break the period down by error rate, cost by model, and the most active agents.

LLM calls counts inference calls: the spans the trace view badges as LLM. Agent, tool, and retrieval spans are not counted, even when they carry a model, and neither are embedding calls, although their cost appears in the cost-by-model breakdown. Token and cost totals, by contrast, cover every span that carries usage, with each model call counted once. A few setups still count a call twice; see Still counted twice.

Token totals include prompt cache reads and writes, whichever way the provider reports them, so totals compare across providers; see Computed server-side.

Latency percentiles (p50, p95, p99) count finished traces only. A run still in progress, such as a long coding session, does not stretch them; it counts once it ends.

Time to first token is how long a streamed LLM call waits before its first chunk arrives: p50 and p95 for the workspace, a value per agent, and a p95 per model in the models table. Only streamed calls record it, so the panel says what share of LLM calls it was measured on. A workspace whose calls never stream shows a dash.

Cost is computed server-side from model pricing and the token usage your spans carry; your instrumentation never ships pricing tables. The view refreshes on the same auto-refresh interval as the trace list, so it doubles as a live wallboard during an agent rollout.

Next steps

Track LLM and agent cost: how each call is priced, what an unpriced model shows, and how to alert on cost.
Exploring traces: drill from a number on this page down to the run behind it.
Alerts: turn a number you keep checking into a notification instead.
Sampling: reduce volume when the numbers get big.