First alert
Connect a notification channel, route an alert to it, and verify the complete delivery path.
This guide verifies the same routing and escalation path used by a real outage. You do not need to break a monitor, and you do not need a monitor at all to run the synthetic test alert.
Before you begin
You need permission to create notification channels and escalation policies. Decide on one destination you can check now, such as your inbox or a test chat channel.
1. Connect a notification channel
- Open Integrations in the workspace sidebar.
- Open Channels and select Add channel.
- Choose a provider and complete the fields shown for that provider.
- Save the channel. If the provider offers a connectivity test, run it now.
Start with one destination. Add fallbacks only after the first path has delivered successfully.
2. Add the channel to an escalation policy
- Open On-call → Escalation policies.
- Edit the workspace default policy, or create a policy for this test.
- Add a first step with a delay of
0minutes. - Choose Channel as the target type and select the channel you just connected.
- Save the policy.
Policies may also target a person, schedule, team, or broadcast. Those targets depend on the recipients' personal notification chains. A direct channel is the easiest first test because it has fewer moving parts.
3. Make the policy the workspace default
A new alert selects its policy in this order: the first matching non-default
routing rule, a policy stored on the monitor, then the workspace default policy. API
clients can set escalation_policy_id when they create or update a monitor. The current
web flow in this guide configures routing rules and the workspace default.
For the smallest setup, use the default policy selector on the escalation policies page. When you later need different paths for services or alert sources, add routing rules and keep the default as a safe fallback.
4. Send a test alert
On On-call → Escalation policies, select Send test alert in the page header. HowlOps creates a synthetic tenant alert and runs it through routing, escalation, and delivery. It resolves automatically after 60 seconds and the action is limited to once every five minutes.
See what happens after a signal arrives.
Choose a source. The first checks change, but routing, paging, response, and recovery stay connected.
Signal
A probe reports that the configured target failed.
Confirm
Retries, thresholds, and optional regional rules confirm the condition before opening an alert.
Route
Routing rules select a policy; without one, enabled channels receive the alert.
Reach
The policy pages its current target and continues only while the alert is unacknowledged.
Respond
Acknowledge, coordinate, and publish updates. Recovery resolves the alert and sends the all-clear.
You should receive the opening alert and its recovery. The alert also appears in the workspace alert list, where you can inspect its routing and delivery timeline.
Test alert or test my on-call?
| Action | Use it to verify |
|---|---|
| Send test alert | Workspace routing, escalation policy selection, and delivery. |
| Test my on-call | Only your personal push → email → SMS → voice chain. |
Use Send test alert for this guide. Use Test my on-call after configuring your personal chain in Settings → Notifications.
If nothing arrives
Do not rebuild the policy immediately. Open the test alert and inspect its delivery timeline, then follow Missing notifications. The most useful first distinction is whether the alert matched no recipient or whether a selected provider failed delivery.
Was this page helpful?