Outgoing Webhook
Learn about outgoing webhooks and how events are triggered for your account.
Last updated
Learn about outgoing webhooks and how events are triggered for your account.
Last updated
Outgoing Webhooks are requests made by PagerTree to a user-defined URL, the payload provided represents an alert. Webhooks allow you to extend PagerTree by performing any logic you wish.
This documentation is meant for webhooks that are outgoing from PagerTree. If you are looking for incoming webhooks, please refer to the incoming webhook documentation.
Create the integration by clicking the Outgoing Webhook logo.
Provide the URL you want PagerTree to send webhooks to.
Outgoing webhooks will send data in a POST request in JSON format to your configured URL. Each request has the following format:
Where event_type
is an applicable event and data
is an alert.
Outgoing webhooks can also send data in a custom format. The request is still a POST request and must be in JSON format.
You can edit the format in the JSON Template section of the integration settings.
Templates support Handlebars substitution with the event
and alert
objects. You can use any handlebars-helpers to support any logic. Use JSON dot notation selection to access deeply nested data (ex: alert.d_user.name
).
The data you have access to are the following properties:
Use the triple-stash {{{ }}} operator to bypass URL encoding.
For the Slack Incoming Webhook App.
If for any reason you do not receive the webhook with a custom template, it is likely you have a formatting error. On the integration page, look at the integration’s logs.
If you do not see a log, ensure the integration is enabled
If you do see a log, view the log by clicking the log link. An error should be provided in the status
field in the content section on the right hand side.
Linked data is a convenience option that embeds referenced objects. Linked data will send the following extra attributes.
Extra Attribute | Referenced By |
---|---|
d_team | d_team_id |
d_user | d_user_id |
source | source_id |
s_log | s_log_id |
If you are using a custom template you can access the linked data like so: {{alert.d_user.name}}
.
Currently PagerTree support 8 events:
alert.created
- fired exactly once, when the alert is created in the database.
alert.open
- Fired exactly once, when the alert transitions from queued -> open.
alert.acknowledged
- fired 0-N times, when the alert is acknowledged by a user.
alert.rejected
- can be fired 0-N times, when the alert is rejected by a user.
alert.timeout
- can be fired 0-M times, when an alert layer times out and moves to the next escalation layer.
alert.resolved
- fired 0-1 times, when the alert is resolved.
alert.dropped
- fired 0-1 times, when the alert is dropped.
alert.handoff
- fired 0-N times, when the alert is handed off.
You can suppress outgoing webhooks or modify the payload using rules. This feature uses the same engine as the routers and is also expressed in YAML. You can use all the same operators as routers for matching conditions, but you can only use the setval and ignore actions. The root element should always be rules
(an array), with each rule having a match
(hash) and actions
(array).
When routers are matching rules they are given access to data. Namely you are given access to the alert
, integration
, and event
in context, and a special field always
which is always true (especially useful for catch all rules).
In our most basic example, we can suppress events when the alert is created by a specific integration.
In a second example, we suppress any event that is not “critical”.
In the next example, we are working with the Slack (Post to Channel) integration and we switch the channel PagerTree sends the notification to.
There is a special value webhook_url
that if set using the setval operator will change the URL that the webhook is sent to. You can use this to dynamically change where webhooks are sent to.
Responses in the 2xx family will be considered a success, any other response will be considered a failure.
Webhooks retries will be attempted up to 14 times with exponential backoff.