Alerts
Problems Spectate or an alert source detected. An alert escalates along its escalation path until someone acknowledges it, and resolves on its own when the problem is over, or when someone resolves it by hand. Resolved alerts can be deleted. Alerts can lead to incidents.
List alerts
Returns the team's alerts, most recently fired first unless you sort.
query Parameters
pagePage number, from 1.
per_pageAlerts per page, 1 to 100.
filter[status]Only firing or resolved alerts.
Whether an alert's problem is ongoing.
firing The problem is ongoing; it escalates until acknowledged. |
resolved The problem is over. |
filter[severity]Only alerts of this severity.
How urgent an alert or incident is. Escalation paths can branch on it.
critical Customers are affected right now. |
high Serious, needs attention soon. |
medium Should be looked at. |
low Informational. |
filter[source_type]Only alerts raised by this kind of check or integration.
What raised an alert.
uptime A monitor went down. |
certificate A monitor's certificate is invalid or expires soon. |
performance A monitor got slow. |
dns The answer of a DNS monitor changed (when it alerts on changes), or a nameserver has a problem that leaves its monitor up. |
host_metric A host's metric condition held, or it stopped sending data. |
domain A domain expires soon, expired or its registration changed. |
external An alert source (external tool) fired it. |
status_page Visitor issue reports over the threshold or a traffic spike ( metadata.type: issue_reports, traffic_spike). |
manual Someone paged by hand, with nothing else to attach it to. |
filter[resource_id]Only alerts about this resource (a monitor, host, domain, status page or alert source ID).
filter[acknowledged]Only acknowledged (true) or unacknowledged (false) alerts.
sortSort by fired_at, resolved_at, severity or status. Prefix - for descending. Default -fired_at.
List alerts › Responses
Paginated set of Alert
current_pagefirst_page_urlfromlast_page_urllast_pageGenerated paginator links.
next_page_urlpathBase path for paginator generated URLs.
per_pageNumber of items shown per page.
prev_page_urltoNumber of the last item in the slice.
totalTotal number of items being paginated.
Get an alert
Returns the alert with the resource it's about and who acknowledged it.
path Parameters
alertAlert ID.
Get an alert › Responses
The alert.
idAlert ID
titleWhat's wrong
descriptionMore about it, as the source described it
statusWhether the problem is ongoing
firing The problem is ongoing; it escalates until acknowledged. |
resolved The problem is over. |
severityHow urgent it is
critical Customers are affected right now. |
high Serious, needs attention soon. |
medium Should be looked at. |
low Informational. |
source_typeWhat raised it
uptime A monitor went down. |
certificate A monitor's certificate is invalid or expires soon. |
performance A monitor got slow. |
dns The answer of a DNS monitor changed (when it alerts on changes), or a nameserver has a problem that leaves its monitor up. |
host_metric A host's metric condition held, or it stopped sending data. |
domain A domain expires soon, expired or its registration changed. |
external An alert source (external tool) fired it. |
status_page Visitor issue reports over the threshold or a traffic spike ( metadata.type: issue_reports, traffic_spike). |
manual Someone paged by hand, with nothing else to attach it to. |
fired_atWhen it fired
acknowledged_atWhen it was acknowledged, which stopped its escalation; null while unacknowledged
acknowledged_byWho acknowledged it: a member's name, an API token, or how (for example a phone call)
recovering_sinceWhen the monitor came back up. The alert stays firing for the team's recovery wait and
resolves when the monitor stays up; null at any other time
group_leader_idThe alert that escalates for this one, when they're grouped; null when it escalates by itself
snoozed_untilUntil when someone snoozed it; null when it isn't snoozed. It counts as acknowledged until then
muted_atWhen maintenance muted it, so it pages nobody until the window ends; null when it isn't muted
resolved_atWhen it resolved; null while firing
resolved_byWho resolved it by hand: a member's name or an API token. null while firing, and when it
resolved on its own because the problem was over
Labels the source sent (Prometheus, Grafana); null when it sent none
Links the source sent along
Delete an alert
Deletes a resolved alert for good, with its escalation history. Incidents and status page incidents it led to stay, without the link to it; messages already posted in Slack or Telegram stay in their channels. A firing alert can't be deleted: resolve it first.
path Parameters
alertAlert ID.
Delete an alert › Responses
No content
Acknowledge an alert
Stops its escalation: nobody else is notified. The alert keeps firing until the problem is over. The dashboard, Slack and Telegram show it as acknowledged by the API token. Acknowledging an acknowledged or resolved alert changes nothing.
path Parameters
alertAlert ID.
Acknowledge an alert › Responses
The acknowledged alert.
idAlert ID
titleWhat's wrong
descriptionMore about it, as the source described it
statusWhether the problem is ongoing
firing The problem is ongoing; it escalates until acknowledged. |
resolved The problem is over. |
severityHow urgent it is
critical Customers are affected right now. |
high Serious, needs attention soon. |
medium Should be looked at. |
low Informational. |
source_typeWhat raised it
uptime A monitor went down. |
certificate A monitor's certificate is invalid or expires soon. |
performance A monitor got slow. |
dns The answer of a DNS monitor changed (when it alerts on changes), or a nameserver has a problem that leaves its monitor up. |
host_metric A host's metric condition held, or it stopped sending data. |
domain A domain expires soon, expired or its registration changed. |
external An alert source (external tool) fired it. |
status_page Visitor issue reports over the threshold or a traffic spike ( metadata.type: issue_reports, traffic_spike). |
manual Someone paged by hand, with nothing else to attach it to. |
fired_atWhen it fired
acknowledged_atWhen it was acknowledged, which stopped its escalation; null while unacknowledged
acknowledged_byWho acknowledged it: a member's name, an API token, or how (for example a phone call)
recovering_sinceWhen the monitor came back up. The alert stays firing for the team's recovery wait and
resolves when the monitor stays up; null at any other time
group_leader_idThe alert that escalates for this one, when they're grouped; null when it escalates by itself
snoozed_untilUntil when someone snoozed it; null when it isn't snoozed. It counts as acknowledged until then
muted_atWhen maintenance muted it, so it pages nobody until the window ends; null when it isn't muted
resolved_atWhen it resolved; null while firing
resolved_byWho resolved it by hand: a member's name or an API token. null while firing, and when it
resolved on its own because the problem was over
Labels the source sent (Prometheus, Grafana); null when it sent none
Links the source sent along
Escalate a muted alert now
Starts its escalation, also while the maintenance window that muted it still runs. An alert that isn't muted doesn't change.
path Parameters
alertAlert ID.
Escalate a muted alert now › Responses
The alert.
idAlert ID
titleWhat's wrong
descriptionMore about it, as the source described it
statusWhether the problem is ongoing
firing The problem is ongoing; it escalates until acknowledged. |
resolved The problem is over. |
severityHow urgent it is
critical Customers are affected right now. |
high Serious, needs attention soon. |
medium Should be looked at. |
low Informational. |
source_typeWhat raised it
uptime A monitor went down. |
certificate A monitor's certificate is invalid or expires soon. |
performance A monitor got slow. |
dns The answer of a DNS monitor changed (when it alerts on changes), or a nameserver has a problem that leaves its monitor up. |
host_metric A host's metric condition held, or it stopped sending data. |
domain A domain expires soon, expired or its registration changed. |
external An alert source (external tool) fired it. |
status_page Visitor issue reports over the threshold or a traffic spike ( metadata.type: issue_reports, traffic_spike). |
manual Someone paged by hand, with nothing else to attach it to. |
fired_atWhen it fired
acknowledged_atWhen it was acknowledged, which stopped its escalation; null while unacknowledged
acknowledged_byWho acknowledged it: a member's name, an API token, or how (for example a phone call)
recovering_sinceWhen the monitor came back up. The alert stays firing for the team's recovery wait and
resolves when the monitor stays up; null at any other time
group_leader_idThe alert that escalates for this one, when they're grouped; null when it escalates by itself
snoozed_untilUntil when someone snoozed it; null when it isn't snoozed. It counts as acknowledged until then
muted_atWhen maintenance muted it, so it pages nobody until the window ends; null when it isn't muted
resolved_atWhen it resolved; null while firing
resolved_byWho resolved it by hand: a member's name or an API token. null while firing, and when it
resolved on its own because the problem was over
Labels the source sent (Prometheus, Grafana); null when it sent none
Links the source sent along
Resolve an alert
Ends the alert by hand, as if its problem were over: its escalation stops, the channels and on-call members it notified hear that it resolved, Slack and Telegram show it as resolved by the API token, and status pages automated from an uptime alert show the monitor as recovered. Use it for false alarms, or for alerts whose source can't tell they're over.
If the problem is still there, the source fires a new alert at its next check (a monitor that is still down, a host metric still over its limit). Resolving a resolved alert changes nothing.
path Parameters
alertAlert ID.
Resolve an alert › Responses
The resolved alert.
idAlert ID
titleWhat's wrong
descriptionMore about it, as the source described it
statusWhether the problem is ongoing
firing The problem is ongoing; it escalates until acknowledged. |
resolved The problem is over. |
severityHow urgent it is
critical Customers are affected right now. |
high Serious, needs attention soon. |
medium Should be looked at. |
low Informational. |
source_typeWhat raised it
uptime A monitor went down. |
certificate A monitor's certificate is invalid or expires soon. |
performance A monitor got slow. |
dns The answer of a DNS monitor changed (when it alerts on changes), or a nameserver has a problem that leaves its monitor up. |
host_metric A host's metric condition held, or it stopped sending data. |
domain A domain expires soon, expired or its registration changed. |
external An alert source (external tool) fired it. |
status_page Visitor issue reports over the threshold or a traffic spike ( metadata.type: issue_reports, traffic_spike). |
manual Someone paged by hand, with nothing else to attach it to. |
fired_atWhen it fired
acknowledged_atWhen it was acknowledged, which stopped its escalation; null while unacknowledged
acknowledged_byWho acknowledged it: a member's name, an API token, or how (for example a phone call)
recovering_sinceWhen the monitor came back up. The alert stays firing for the team's recovery wait and
resolves when the monitor stays up; null at any other time
group_leader_idThe alert that escalates for this one, when they're grouped; null when it escalates by itself
snoozed_untilUntil when someone snoozed it; null when it isn't snoozed. It counts as acknowledged until then
muted_atWhen maintenance muted it, so it pages nobody until the window ends; null when it isn't muted
resolved_atWhen it resolved; null while firing
resolved_byWho resolved it by hand: a member's name or an API token. null while firing, and when it
resolved on its own because the problem was over
Labels the source sent (Prometheus, Grafana); null when it sent none
Links the source sent along
Page someone
With nothing to attach it to: creates an alert with source manual, open until someone
resolves it or its incident resolves. With an incident_id whose alert still fires, the page
goes on that alert instead. Texts and calls use the team's credits.
Page someone › Request Body
target_typeWho: a member, whoever is on call for a schedule, or an escalation path
member A team member. |
schedule Whoever is on call for a schedule. |
path An escalation path. |
target_idThe member's, schedule's or path's ID
titleWhat the page is about, up to 255 characters
messageMore detail for the person paged, up to 2000 characters
severityHow urgent. Default high
critical Customers are affected right now. |
high Serious, needs attention soon. |
medium Should be looked at. |
low Informational. |
incident_idAn incident to tie this to: the new alert resolves with it
Page someone › Responses
The alert the page is on.
idAlert ID
titleWhat's wrong
descriptionMore about it, as the source described it
statusWhether the problem is ongoing
firing The problem is ongoing; it escalates until acknowledged. |
resolved The problem is over. |
severityHow urgent it is
critical Customers are affected right now. |
high Serious, needs attention soon. |
medium Should be looked at. |
low Informational. |
source_typeWhat raised it
uptime A monitor went down. |
certificate A monitor's certificate is invalid or expires soon. |
performance A monitor got slow. |
dns The answer of a DNS monitor changed (when it alerts on changes), or a nameserver has a problem that leaves its monitor up. |
host_metric A host's metric condition held, or it stopped sending data. |
domain A domain expires soon, expired or its registration changed. |
external An alert source (external tool) fired it. |
status_page Visitor issue reports over the threshold or a traffic spike ( metadata.type: issue_reports, traffic_spike). |
manual Someone paged by hand, with nothing else to attach it to. |
fired_atWhen it fired
acknowledged_atWhen it was acknowledged, which stopped its escalation; null while unacknowledged
acknowledged_byWho acknowledged it: a member's name, an API token, or how (for example a phone call)
recovering_sinceWhen the monitor came back up. The alert stays firing for the team's recovery wait and
resolves when the monitor stays up; null at any other time
group_leader_idThe alert that escalates for this one, when they're grouped; null when it escalates by itself
snoozed_untilUntil when someone snoozed it; null when it isn't snoozed. It counts as acknowledged until then
muted_atWhen maintenance muted it, so it pages nobody until the window ends; null when it isn't muted
resolved_atWhen it resolved; null while firing
resolved_byWho resolved it by hand: a member's name or an API token. null while firing, and when it
resolved on its own because the problem was over
Labels the source sent (Prometheus, Grafana); null when it sent none
Links the source sent along
Page someone about an alert
Pages a team member, whoever is on call for a schedule, or starts an escalation path, also if
the alert is acknowledged. Texts and calls use the team's credits. The target must belong to
the team (422 otherwise).
path Parameters
alertAlert ID.
Page someone about an alert › Request Body
target_typeWho: a member, whoever is on call for a schedule, or an escalation path
member A team member. |
schedule Whoever is on call for a schedule. |
path An escalation path. |
target_idThe member's, schedule's or path's ID
Page someone about an alert › Responses
The alert.
idAlert ID
titleWhat's wrong
descriptionMore about it, as the source described it
statusWhether the problem is ongoing
firing The problem is ongoing; it escalates until acknowledged. |
resolved The problem is over. |
severityHow urgent it is
critical Customers are affected right now. |
high Serious, needs attention soon. |
medium Should be looked at. |
low Informational. |
source_typeWhat raised it
uptime A monitor went down. |
certificate A monitor's certificate is invalid or expires soon. |
performance A monitor got slow. |
dns The answer of a DNS monitor changed (when it alerts on changes), or a nameserver has a problem that leaves its monitor up. |
host_metric A host's metric condition held, or it stopped sending data. |
domain A domain expires soon, expired or its registration changed. |
external An alert source (external tool) fired it. |
status_page Visitor issue reports over the threshold or a traffic spike ( metadata.type: issue_reports, traffic_spike). |
manual Someone paged by hand, with nothing else to attach it to. |
fired_atWhen it fired
acknowledged_atWhen it was acknowledged, which stopped its escalation; null while unacknowledged
acknowledged_byWho acknowledged it: a member's name, an API token, or how (for example a phone call)
recovering_sinceWhen the monitor came back up. The alert stays firing for the team's recovery wait and
resolves when the monitor stays up; null at any other time
group_leader_idThe alert that escalates for this one, when they're grouped; null when it escalates by itself
snoozed_untilUntil when someone snoozed it; null when it isn't snoozed. It counts as acknowledged until then
muted_atWhen maintenance muted it, so it pages nobody until the window ends; null when it isn't muted
resolved_atWhen it resolved; null while firing
resolved_byWho resolved it by hand: a member's name or an API token. null while firing, and when it
resolved on its own because the problem was over
Labels the source sent (Prometheus, Grafana); null when it sent none
Links the source sent along
Snooze an alert
Pauses its escalation for 15 minutes to 24 hours. It counts as acknowledged until then; if it
still fires then, the escalation starts again from the first step. Resolving ends it. A
resolved alert can't be snoozed (422).
path Parameters
alertAlert ID.
Snooze an alert › Request Body
minutesHow long in minutes, 15 to 1440
reasonWhy, for the team; up to 500 characters
Snooze an alert › Responses
The snoozed alert.
idAlert ID
titleWhat's wrong
descriptionMore about it, as the source described it
statusWhether the problem is ongoing
firing The problem is ongoing; it escalates until acknowledged. |
resolved The problem is over. |
severityHow urgent it is
critical Customers are affected right now. |
high Serious, needs attention soon. |
medium Should be looked at. |
low Informational. |
source_typeWhat raised it
uptime A monitor went down. |
certificate A monitor's certificate is invalid or expires soon. |
performance A monitor got slow. |
dns The answer of a DNS monitor changed (when it alerts on changes), or a nameserver has a problem that leaves its monitor up. |
host_metric A host's metric condition held, or it stopped sending data. |
domain A domain expires soon, expired or its registration changed. |
external An alert source (external tool) fired it. |
status_page Visitor issue reports over the threshold or a traffic spike ( metadata.type: issue_reports, traffic_spike). |
manual Someone paged by hand, with nothing else to attach it to. |
fired_atWhen it fired
acknowledged_atWhen it was acknowledged, which stopped its escalation; null while unacknowledged
acknowledged_byWho acknowledged it: a member's name, an API token, or how (for example a phone call)
recovering_sinceWhen the monitor came back up. The alert stays firing for the team's recovery wait and
resolves when the monitor stays up; null at any other time
group_leader_idThe alert that escalates for this one, when they're grouped; null when it escalates by itself
snoozed_untilUntil when someone snoozed it; null when it isn't snoozed. It counts as acknowledged until then
muted_atWhen maintenance muted it, so it pages nobody until the window ends; null when it isn't muted
resolved_atWhen it resolved; null while firing
resolved_byWho resolved it by hand: a member's name or an API token. null while firing, and when it
resolved on its own because the problem was over
Labels the source sent (Prometheus, Grafana); null when it sent none
Links the source sent along