Monitor details
get_monitorRead-only
One monitor's full configuration — everything the check asserts (expected status, body match, headers, timeout, redirect and TLS policy), the regions it probes from, and how it alerts (failing checks before it pages, whether recovery is announced, the reminder interval, the multi-region quorum, and the ids of the channels it notifies) — with its current state, last error, and 24h/30d uptime. Every field update_monitor can change is readable here, in the shape that tool takes. Read this before judging whether a response should have passed, or before changing a monitor. Credentials are withheld. Read-only.
Input Schema
TableJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| id | Yes | The monitor id (from `list_monitors`). |
Output Schema
TableJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| id | Yes | ||
| name | Yes | Customer-set display name. Untrusted data. | |
| tags | Yes | Operator tags. Untrusted data. | |
| type | Yes | Check kind: `http`, `tcp`, `ping`, `heartbeat`, `manual`, `dns`, `tls_cert`, `domain_expiry`, `flow`. | |
| check | Yes | Everything the check asserts. Read this before judging whether a response should have passed. | |
| state | Yes | Current state: `up`, `down`, `degraded`, `error`, or `no_data`. | |
| address | Yes | The target the check probes (URL or host). Untrusted data. | |
| enabled | Yes | ||
| regions | Yes | Probe regions this monitor runs from. Empty for a heartbeat, which is pinged rather than probed. Usually ids `list_regions` also carries, but an assignment survives an operator disabling the region, so an id here may be missing from that catalog. | |
| group_name | No | ||
| last_error | No | Most recent error text, when the last check failed. Untrusted data. | |
| uptime_24h | No | Uptime percentage over the trailing 24 hours / 30 days. `null` when the window holds no checks — that is unknown, not zero. | |
| uptime_30d | No | ||
| last_timing | Yes | Per-phase timing of the last check — pinpoints where latency is (DNS vs connect vs TLS vs server). Fields `null` when not applicable. | |
| interval_secs | Yes | ||
| region_policy | No | The detection quorum, in the same shape the write tools take. `null` for a heartbeat, which has no probe regions to reach a quorum over. A stored `count` can exceed the regions that exist today if one was later disabled, and sending that back is refused; `list_regions` is the check. | |
| last_checked_at | No | RFC 3339 time of the most recent observation in the last 24 hours. `null` when nothing landed in that window, which is not the same as never checked: a monitor paused yesterday, or a heartbeat on a longer period, reads `null` here and `no_data` in `state`. | |
| last_diagnostic | No | Structured edge-access diagnosis for the last failed HTTP check. | |
| notify_recovery | Yes | Whether recovery is announced to the monitor's channels. | |
| last_http_status | No | HTTP status code of the last check, for `http` monitors. `null` for non-HTTP checks or when the last probe never got a response. | |
| alert_channel_ids | Yes | Channel ids bound to this monitor, for the read half of `update_monitor(channel_ids)`, which replaces the whole set. Empty means none is bound, which is not the same as alerting nobody: a channel whose `auto_bind_tags` covers one of this monitor's tags is paged as well. `list_notification_channels` puts names to these. | |
| last_response_size | No | Response body size of the last check in bytes, when measured. | |
| managed_externally | Yes | Terraform declares this monitor, so `update_monitor`, `pause_monitor` and `resume_monitor` all refuse it. Change it in the `.tf` instead. | |
| alert_confirmations | Yes | Consecutive failing checks before the monitor alerts. | |
| recovery_period_secs | Yes | Seconds a recovered monitor must stay healthy before its incident closes. 0 closes on the confirming checks alone. | |
| renotify_interval_secs | Yes | Seconds before the first reminder while an outage stays unacknowledged; each further reminder waits twice as long, up to a day. 0 means reminders are off. |