Skip to Content

How can my coding agent list firing alerts in Rius?

Your coding agent lists firing alerts in Rius with the list_alerts tool on the Rius MCP server. Ask “What alerts are open right now?” and the agent calls it with status set to open. The reply is a Markdown list of alerts, newest first, that your client renders as a table.

How it works

An alert is opened by a detector and stays open, collecting occurrences, until the detector resolves it. An open alert with a recent last_seen is still happening, so “firing” in plain words means status of open.

You rarely call the tool by hand. Your client reads your question and picks the tool and arguments. These are the main parameters it can use:

  • status: open or resolved. Omit it to get both.
  • kind: one detector, either consistent_error_pattern, swallowed_error, tool_loop or custom. Every custom rule shares the kind custom, so use rule_id to isolate a single custom rule.
  • rule_id: the firings of one rule, taken from list_alert_rules.
  • service and step: the exact service name or step (span) name.
  • search: a case-insensitive match on the alert title.
  • acknowledged: true or false, to separate alerts someone has seen from those nobody has.
  • hours: how far back to look, measured on last_seen. The default is 24, and “this week” is 168.
  • limit: rows to return, 50 by default and at most 200.
  • workspace: a workspace name or ID. Omit it to use your default workspace.

An alert last seen before the window falls outside it, so widen hours when you expect older alerts. The tool reference lists every argument.

Set it up

  1. Connect your client to the Rius MCP server. In Claude Code, run this in your terminal and complete the browser sign-in the first time a session uses the server:

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

    If you already use the Rius plugin, it registers the server as rius, so you do not add it again. Other clients, and agents that run without a browser and use a read-scoped API key, are covered on the MCP page.

  2. Ask your agent in plain words. These questions map to the tool:

    • “What alerts are open right now?”
    • “Show me the errors alert for the checkout service this week.”
    • “Which tool loop alerts have nobody acknowledged?”

  3. To see what the client sends for the first question:

    { "name": "list_alerts", "arguments": { "status": "open" } }

  4. For a service and a longer window, the client sends:

    { "name": "list_alerts", "arguments": { "service": "checkout", "hours": 168 } }

  5. Follow up on one alert. Ask “Why did that alert fire?” and the agent passes the alert’s id from the list to get_alert. That call returns the alert’s labels, the sample trace ids behind it and, if root-cause analysis ran, the written finding. Pass a sample trace id to get_agent_trace for the span waterfall.

What you get

  • A list of fired alerts, newest first, that you can narrow by status, detector, rule, service, step, title text and acknowledgement.
  • Open alerts that are still accumulating occurrences, so your agent can tell “still happening” from “already resolved”.
  • Alert ids you can hand to get_alert for the full detail and root-cause finding.
  • A path from a firing alert to the traces behind it, inside one coding session. For the wider picture of watching agents in production, see Monitor AI agents in production.
  • Read access with an OAuth session or a read-scoped API key. The alert reads are among the tools a key without manage scopes can reach, and the key is bound to the one workspace it was created in.

When something else fits

  • If you want to stop notifications for an alert, ack_alert records that a person has seen it. It works only on an interactive OAuth connection, because an API key is not a user, and the alert stays open until the detector resolves it.
  • If you want to change which alerts exist, use the alert rule and channel tools, which need a key minted with alerts:manage or channels:manage, or an admin’s OAuth session. See the MCP page for the access rules.
  • If you want to browse alerts yourself rather than ask an agent, the Rius console shows the same firings.

Still stuck? Get help

Last updated on
Quickstart