Skip to main content
Vigil is an open-source microservices status page and probing system. With Vigil’s webhook notifier ([notify.webhook]), you can sync the overall status of the status page to Flashduty On-call: an alert is triggered when the status becomes dead or sick, and it recovers automatically when the status returns to healthy. Vigil’s webhook body has a fixed format, so no template is needed.

​
In Flashduty On-call

You can get the integration push URL in either of the following ways.

​
Use a dedicated integration

  1. Go to the Flashduty console, select Channels, and open a channel
  2. Select Settings → Integrations → Dedicated integrations, and click Add an integration
  3. Select Vigil and click Save
  4. Open the generated integration card and copy the Push URL

​
Use a shared integration

  1. Go to the Flashduty console and select Integration Center → Alert Events
  2. Select Vigil and enter an integration name
  3. Configure the default route and select a channel; you can add more rules under Routes after creation
  4. Click Save and copy the generated Push URL

​
In Vigil

You need Vigil 1.29.x and access to edit Vigil’s configuration file (config.cfg). The server running Vigil must be able to reach the Flashduty push URL.
1

Configure the webhook notifier

In config.cfg, add (or edit) [notify.webhook] and set hook_url to the full Flashduty push URL, which must include integration_key:
Flashduty authenticates with the integration_key in the URL, so no extra headers are needed. reminder_interval is how often Vigil repeats the notification while the status stays abnormal, in seconds; adjust it as needed. Restart Vigil after saving so the configuration takes effect.
2

Configure probe targets

Under [[probe.service]], configure the services and nodes to monitor. Vigil combines the status of all nodes into the overall status of the status page. The notifier needs no other settings.
3

Verify

Vigil has no test button. To verify:
  1. Make a probed node fail, wait for Vigil to mark it as dead, and confirm a Critical alert appears in Flashduty with the failed node listed in its description
  2. Let the node recover and confirm the alert recovers automatically in Flashduty (see the note below on sick)
If startup_notification = true is set under [notify], Vigil sends a startup notification when it starts. Flashduty returns success without creating an alert, which you can use to confirm the URL and network connectivity.

​
Alert Key

Flashduty uses the status page URL page.url as the Alert Key. Vigil notifications describe only the aggregate status of the whole status page and carry no per-alert ID, so one Vigil status page maps to one Flashduty alert: it is created or updated when the status becomes dead or sick, and recovered when it becomes healthy.
  • When several nodes fail at once, one notification lists all failing nodes. If one node recovers while another is still failing, the status stays dead, the alert stays open, and the node list in its description shrinks until the overall status is healthy
  • The recovery notification has an empty replicas list, so it cannot say which node recovered
  • Changes to the page title, status, node list, or notification time do not change the Alert Key. A request without page.url is rejected
  • page.url comes from page_url under [branding] in config.cfg and is unique only within one Vigil instance. Do not give several Vigil instances the same page_url, and send only one Vigil instance to each Flashduty integration

​
Status and severity

Vigil sends a notification only when the status becomes dead, when it leaves dead (to sick or healthy), and as reminders while it stays dead. A change from healthy to sick, or from sick to healthy, sends nothing. A sick notification therefore only follows dead, always carries an empty replicas list, and updates the same alert to Warning. If the page then goes from sick to healthy, Vigil sends no recovery, so the alert stays open until the page turns dead and healthy again, or until you close it in Flashduty. Any other status is rejected. Vigil’s notification time has only hours, minutes and seconds with no date, so Flashduty uses the time it receives the request as the event time. The alert title looks like Vigil: Example Status is dead, built from the status page title (or the page URL when the title is empty) and the status. The description lists the failing nodes in alphabetical order, in the format service ID:node ID:probe URL.

​
Labels

​
Troubleshooting

  • Flashduty returns an invalid-parameter error: confirm hook_url is complete and includes integration_key; the response names the missing field or the unsupported status
  • No alert received: confirm Vigil was restarted, hook_url under [notify.webhook] is correct, and the overall status Vigil computes actually became dead or sick; Vigil sends notifications only on status changes and when the reminder interval elapses
  • Alert does not recover: a recovery is sent only when the whole status page goes from dead straight to healthy; while any node is still failing the status stays dead or sick, and a sick to healthy change is not notified, so close such an alert manually
  • Repeated notifications for the same status page: these are the reminders triggered by reminder_interval, and Flashduty merges them into the same alert
For more details, see the Vigil project page.