How can my coding agent list the workspace metrics in Rius?

Rius lets your coding agent list the workspace metrics with the list_metrics_catalog tool on the Rius MCP server. Ask in plain words, such as “What latency metrics exist?”, and your client calls the tool and returns the metrics as a Markdown table. Narrow the list with the optional group and kind arguments. Nothing else is required.

How it works

list_metrics_catalog returns every metric Rius knows how to compute. For each metric it gives:

• the unit
• whether the value can be null
• which read serves it
• its companions, the extra response keys the metric contributes. Scalars add coverage keys, and breakdowns and series add row or point value keys

Use it to find exact field names instead of guessing them. The reply is Markdown, which your client renders as a table.

The tool takes three optional arguments:

You rarely call the tool by hand. Your client reads your question and picks the tool and arguments itself. Questions that map to it include:

• “List the workspace metrics in Rius.”
• “What cost metrics can I read?”
• “Which metrics come as time series?”
• “What latency metrics exist?”

The last question sends this call:

{ "name": "list_metrics_catalog", "arguments": { "group": "latency" } }

The tool is read-only, so it changes nothing in your workspace. The full argument list is in the tool reference.

Set it up

1. Connect your coding agent to the Rius MCP server. In Claude Code, run this in your terminal. The OAuth sign-in opens the first time a session uses the server:

   claude mcp add --transport http rius https://mcp.eu.console.rius-glassflow.com/mcp

   For Codex, run codex mcp add rius --url https://mcp.eu.console.rius-glassflow.com/mcp. Keep --url before the URL. If you already use the Rius plugin for Claude Code, Codex (beta) or Cursor (beta), it registers the server as rius, so you can skip this step. Other clients are covered on the MCP page.

2. Check the connection. Ask your agent “Who am I in Rius, and which workspaces can I access?” It should call get_me and answer with your email and your workspaces.

3. Ask for the catalog in plain words, for example:

   What cost metrics exist in my Rius workspace?

   Your agent sends list_metrics_catalog with group set to cost.

4. To see only one kind of metric, say so in your question, for example “Which metrics are breakdowns?” Your agent sets kind to breakdown.

5. For a headless agent, such as one in CI, use a read-scoped API key sent as an Authorization Bearer header. A key without manage scopes reaches the metrics tools. Ingest keys, the kind the SDK uses, cannot query through the MCP server.

What you get

• A table of the metrics Rius can compute, each with its unit, whether it can be null, the read that serves it, and its companion keys.
• The exact metric and field names, so you stop guessing them.
• A list narrowed to one group (cost, volume, tokens, latency, ttft or reliability) or one kind (scalar, series or breakdown).
• A clear next step. The catalog names the metrics, and workspace_metrics_overview returns their values. That tool gives workspace health over a time window: volume, cost, tokens, latency percentiles, time to first token and error rates. Each comes with its previous-window baseline and delta, plus ranked breakdowns per model, per agent and per kind. Ask “How is my workspace doing today?” to call it.

Both tools read your workspace telemetry. For the wider picture of watching agents in production, see Monitor AI agents in production.

When something else fits

• To read the numbers rather than the list of metrics, ask for workspace_metrics_overview. For one agent or one slice of traffic, agent_traces_summary gives the same kind of figures with filters.
• To search individual traces or spans by field, use describe_search_catalog first. It lists the fields you can filter on, which is a different list from the metrics catalog.
• To browse the same numbers visually, use the Rius console instead of your coding agent.

Related