Webhook alerts still being sent for disabled metrics

[revilo951]

Describe the bug
Webhook alerts still being sent for disabled metrics

To Reproduce
Steps to reproduce the behavior:

1. Disable alert for metric

3. Increase usage on metric above global default threshold
4. Get webhook alert

Expected behavior
No alert for disabled metrics

Environment:

• Pulse Version: 4.23.0
• Installation Type: ProxmoxVE LXC

[rcourtman]

Update: Disabled metrics now clear pending notifications so webhooks stop firing. Commit c6c0ac6.

[revilo951]

Just wondering if this fix is planned to be released some time soon?

And thanks for an amazing bit of software! 😄

[rcourtman]

Already released in v4.24.0 and v4.25.0. Upgrade to get the fix.

[revilo951]

Deleted last comment, I am still seeing an issue where higher thresholds than the global are triggering notifications (which I tried to stop the notifications with this bug). I'll double check.

[revilo951]

Looks like it's still occurring:

Edit: I am an idiot. Disregard this. I will continue testing.

[revilo951]

OK, I'm not going mad after all, it's still happening.

I have noticed that when the page loads, the disabled metric briefly shows "80%" before flicking back to "Off", not sure if that's contributing?

[rcourtman]

Need your Pulse version. Disabled-metric webhook issue is fixed in v4.24.0 and v4.25.0. Migration resets on custom thresholds are addressed by commit 4ae426c but that fix has not shipped in a release yet. If you are running a release build, thresholds will still revert after a VM migration until the next release is out; building from main now will include the change. Confirm whether a migration happened before the notification and include the metric name, alert payload, and service logs for that timestamp if you are already on v4.24.0 or newer.

[revilo951]

Version: 4.25.0

No migrations occuring.

Metric name is frigate

Payload:

Pulse Alert: Warning
Storage at 89.3%
Resource frigate
Node cluster
Type Usage
Value 89.3%
Threshold 80%
Duration 9m
Pulse Monitoring•Today at 07:14

Is this what you need for the service logs?
pulse-diagnostics-sanitized-2025-11-06.json

[rcourtman]

The Oct 16 commit (c6c0ac6) that shipped in v4.24.0 only clears alerts when a per-resource override disables a metric. With a DisableAll* flag, the Check* path bails out before cleanup, so those alerts sit in m.activeAlerts and the webhook loop keeps sending them. I pushed a patch that makes reevaluateActiveAlertsLocked honor every DisableAll flag and resolves the matching alerts, and the Check* methods now drop them before returning when a global disable is active. Added regression coverage for the storage toggle. Pull current main (it ships in the next point release) and the webhook spam stops once the flag flips.

[revilo951]

Fixed! Awesome work, thanks @rcourtman :)