Create a monitor
create_monitorCreates a monitor in a free slot of the workspace's plan. Premium monitors (multi-step transactions and code repositories) have slots of their own, on Pro and Business only. When every slot is in use the call fails with monitor_limit_reached: the user can upgrade at https://upheron.com/pricing or delete a monitor. Needs a write token.
Tool call template
Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.
{
"id": 1,
"jsonrpc": "2.0",
"method": "tools/call",
"params": {
"arguments": {
"name": "<name>"
},
"name": "create_monitor"
}
}Input Schema
| Name | Required | Description | Default |
|---|---|---|---|
| url | No | URL, or host[:port] for network checks. | |
| kind | No | What to check. `heartbeat` and `webhook` wait for your calls instead of probing. `repository` reads a GitHub or GitLab repository's dependency manifests daily and alerts on vulnerabilities in them, on every plan; put the repository URL in `url` (optionally …/tree/branch/folder) and, for a private repository (Pro plan), an access token in `check_options.token`. `websocket` (a ws:// or wss:// URL) and `grpc` (host[:port]) are created on Pro and Business (Free answers `422` with `plan_required`). | http |
| name | Yes | Shown in alerts and on status pages. | |
| charts | No | Charts on public status pages: uptime bars (the last 90 days, or the whole history), the last 24 hours of response time, or both. | both |
| method | No | The HTTP method of the check's request. | GET |
| tab_id | No | The monitor tab (GET /monitor-tabs) the monitor is sorted into in the dashboard; `null` takes it out of its tab. | |
| keyword | No | Text the response body must contain. | |
| dns_watch | No | Every plan, `http` and `api` monitors: look up the host's A, AAAA and CNAME records and its zone's NS, MX, SPF and DMARC hourly, and alert when one changes, with the old and new values. | |
| locations | No | Location codes from GET /locations. Defaults to the first location. | |
| port_scan | No | Business plans, network and website checks: scan the common TCP ports of the host's verified addresses weekly (POST /monitors/{id}/ports/verify), and alert when a port opens. See /docs/port-scans. | |
| ip_version | No | Network and website checks: `auto` connects over IPv4, or IPv6 when the host has no IPv4 address; `ipv4` and `ipv6` use only that. `ipv6` needs locations with IPv6 (`ipv6` in GET /locations). An IP address as the host sets it. | auto |
| slo_target | No | Every plan: an availability objective in percent, such as 99.9, over slo_window_days; null for none. Downtime is the time in incidents, less maintenance. The monitor's contacts are alerted when its error budget burns fast, is 50%, 75% and 90% used, and when it runs out. See /docs/reports#error-budgets. | |
| timeout_ms | No | Milliseconds a check waits for the reply before it counts as failed. | |
| status_page | No | Creating only: give the monitor its own published status page while the plan allows more pages. Send `false` when you manage status pages yourself. | |
| header_watch | No | Every plan, `http` and `api` monitors: grade the page's security headers, cookie flags and HTTPS redirect with each website scan (weekly on Free, daily on paid plans), and alert when one that was there is gone or weaker. See /docs/website-changes. | |
| request_body | No | The body sent with the check's request, such as JSON for a POST. | |
| check_options | No | Kind-specific options, such as `json_path`/`expected_value` for `api`, `steps` for transactions, `warn_days` for ssl. Ports go in `url` as host:port. See /docs/api#check-options. | |
| content_watch | No | Business plans, `http` and `api` monitors: look at the page hourly, keeping only a fingerprint of its text, title and external scripts, and alert when the text changes by `content_threshold` percent or more, the title changes, a script from a new host appears or defacement phrases show up. Accept a change with POST /monitors/{id}/changes/accept. | |
| expiry_alerts | No | `http` and `api` monitors: check the domain registration and (for HTTPS) the TLS certificate daily. Domain alerts come 30, 14, 7, 3 and 1 days before expiry; certificate alerts only once automatic renewal should have happened (a sixth of the certificate's lifetime, at most 14 days). Uses no extra monitor. | |
| repository_url | No | Pro and Business, `http` and `api` monitors with vulnerability alerts: the website's GitHub or GitLab repository (optionally …/tree/branch/folder), whose lockfiles and manifests are read hourly for exact versions. A failed read opens no incident. `null` unlinks it and removes what it listed. | |
| idempotency_key | No | Any unique string. Repeating a call with the same key within 24 hours returns the first monitor instead of creating another. Use one per intended monitor. | |
| request_headers | No | Sent with the check; stored encrypted and never returned. | |
| slo_window_days | No | The rolling window of the availability objective, in days. | |
| interval_seconds | No | Seconds between checks, no faster than the plan allows (300 on Free, 10 on Pro). | |
| lookalike_alerts | No | Pro and Business, `http` and `api` monitors: look up a few hundred typo and homoglyph spellings of the registered domain daily, and alert when one starts resolving or gains MX records. | |
| repository_token | No | Access token for a private `repository_url`: on GitHub a fine-grained token with read-only Contents, on GitLab `read_api`. Stored encrypted and never returned (`repository_token_set` tells whether one is saved). Leave it out to keep it; `null` removes it. | |
| subdomain_alerts | No | Pro and Business, `http` and `api` monitors: list the domain's subdomains from certificate transparency logs (through crt.sh, which can be days or weeks behind), resolve each daily, and alert on a new one and on a CNAME to a service that reports no site there (a subdomain takeover risk). | |
| alert_channel_ids | No | Notification contacts to alert. Defaults to all of the workspace's contacts. | |
| content_threshold | No | With `content_watch`: the share of the page's text, in percent, that must change to alert. Smaller changes become the new baseline. | |
| failure_threshold | No | Failed checks in a row before an incident opens. The default alerts on the first failed check. | |
| recovery_threshold | No | Successful checks in a row before the incident resolves. | |
| alert_delay_minutes | No | Pro and Business plans: minutes the first outage alert waits after the incident opens (on top of `failure_threshold`). Nobody is alerted about, or hears the recovery of, an outage that ends sooner. | |
| dns_address_changes | No | What A and AAAA answers alert on: `network` when they move to another network (autonomous system), so CDNs rotating addresses do not alert; `address` on any address change, for a site on fixed addresses; `none` never. | network |
| expected_status_max | No | The highest HTTP status that counts as up; at least `expected_status_min`. | |
| expected_status_min | No | The lowest HTTP status that counts as up. | |
| escalation_policy_id | No | Business plans: alert this escalation policy (GET /escalation-policies) about outages instead of `alert_channel_ids`, which still receive expiry and vulnerability alerts. `null` alerts the contacts. | |
| repeat_alert_minutes | No | Pro and Business plans: repeat the outage alert to the monitor's contacts this often until someone acknowledges the incident or it recovers, at most 48 times. `null` sends it once. | |
| vulnerability_alerts | No | Every plan, `http` and `api` monitors: scan the website for the software it runs (weekly on Free, daily on paid plans) and alert on newly published vulnerabilities in it, checked daily and within 15 minutes of a new advisory (OSV.dev with the GitHub Advisory Database, NVD, new CVE records, WPVulnerability and WordPress core advisories for WordPress, CISA KEV, endoflife.date). | |
| confirmation_threshold | No | Locations that must fail a check at the same time for it to count as failed. 1, the default, is any location; with 2 of 3 locations, one failing location alone is no outage and alerts nobody. More than the monitor's locations counts as all of them. | |
| vulnerability_severity | No | The lowest severity that alerts. Actively exploited vulnerabilities and releases reaching end of support always alert. | high |
| heartbeat_grace_seconds | No | For heartbeat and webhook monitors. |