Monitor agent

Get the agent

Get a monitor's agent: its status, its plan, what it found, its ticks, its spend and budget, and how each source did. A tick is one run of the agent.

GET/v1/monitors/{monitor_id}/agent

Send your API key as Authorization: Bearer. See Authentication. How each option works: Monitors.

Path parameter

200The monitor's agent. A tick is one run of the agent.

monitor_idstring: The ID of the monitor.
statusstring: The agent's status: compiling, ready, failed, paused, ended or deleted.
querystring: The monitor's query.
frequencystring: How often the agent runs.
spec_versioninteger: The version of the agent's plan. It goes up each time the plan changes.
specobject or null: The plan the agent follows: the kind of object it looks for, the criteria each one must meet, and where it searches. null before the first plan.
countsobject: How many objects the agent found, by status, and its alerts in the last 24 hours.
ticks[]object[]: The agent's latest ticks: when each ran, what it cost, why it stopped and what it found.
next_tick_atstring or null: When the agent runs next, or null.
spend_usdnumber: What the agent has spent on this monitor, in US dollars.
budgetobject or null: The monitor's own budget, as set with Update the agent, or null.
source_health[]object[]: How each source of the plan did: its rows, new objects, cost and status.
compileobject: The agent's latest plan job: its job_id, status and error. See Get a job.
running_jobobject or null: The job that runs now, or null.
suggested_prompts[]string[]: Messages you can send the agent to start a chat.

Status	Code	When
`401`	`unauthorized`	The API key is missing or not valid.
`403`	`forbidden`	The API key cannot read monitors (a `GET`) or change them (the other methods).
`404`	`not_found`	The monitor is not in your key's project, or it has no agent.
`429`	`rate_limited`	Your key or account sent too many requests. Wait for the seconds in `Retry-After`, then try again.

What each error means and what to do: Errors.

Request

curl -X GET "https://router.telem.ai/v1/monitors/$MONITOR_ID/agent" \
  -H "Authorization: Bearer $TELEM_API_KEY"

agent = client.monitors.agent(monitor_id)
print(agent.status, agent.spec_version)

const agent = await telem.monitors.agent(monitorId)
console.log(agent.status, agent.specVersion)

Response

{
  "error": {
    "code": "not_found",
    "message": "This monitor has no agent"
  }
}

Built from the error text that Telem sends.

Was this page helpful?