Auto-disable thresholds
When a webhook fails N times in a row, Webhooks Pro can automatically flip it to
AUTO-DISABLED(red lozenge). The counter resets on the first successful delivery. Default: off. Tenant-wide setting.
The default is off because aggressive auto-disable does more harm than good on flaky receivers. Turn it on only when you've measured your real failure pattern.
Where to set it
Webhooks Pro → Settings → Webhooks → Auto-disable on failure.
Option | Behaviour |
|---|---|
Disabled (default) | Never auto-disable. Failures are just logged. |
3 failures | Disable after 3 consecutive failures. |
5 failures | Disable after 5 consecutive failures. |
10 failures | Disable after 10 consecutive failures. |
20 failures | Disable after 20 consecutive failures. |
The threshold is tenant-wide — applies to every webhook on this site. There is no per-webhook override today.
What counts as a failure
Result | Counted? |
|---|---|
2xx Success | ✅ Reset counter to 0 |
3xx Redirect | ✅ Reset counter to 0 |
4xx Client Error | ❌ Failure — counter ticks |
5xx Server Error | ❌ Failure — counter ticks |
Network Error / timeout (status 0) | ❌ Failure — counter ticks |
Skipped — license | Not counted at all |
Two important corollaries:
One success resets the counter. A 5-out-of-6 failure rate (with one success in the middle) does not trip a threshold of 5.
Licence-skipped deliveries don't count. A tenant-wide licence problem doesn't auto-disable individual webhooks.
When auto-disable fires
delivery #1 → 5xx → counter = 1
delivery #2 → 5xx → counter = 2
delivery #3 → 200 → counter = 0 ← reset
delivery #4 → 5xx → counter = 1
delivery #5 → 5xx → counter = 2
delivery #6 → 5xx → counter = 3 ← if threshold = 3, AUTO-DISABLED hereAfter auto-disable:
Status lozenge on the Webhooks list flips to
AUTO-DISABLED(red).Webhook stops firing on new events.
Configuration is preserved; nothing is changed except the active state.
An admin must re-enable manually.
Re-enable playbook: Re-enable an auto-disabled webhook.
Picking the right threshold
A few rules:
Don't pick
3for production. Three failures from a flaky receiver in 60 seconds is normal; three failures shouldn't kill the integration.5is a good "tight" choice for receivers that are usually reliable.10or20is the right call for receivers known to be flaky during specific times (e.g. during nightly batch). The counter still resets on the first success, so a rough patch isn't enough by itself.If you're not sure, leave it off. Use Webhook logs and external monitoring to alert on failure patterns instead. Auto-disable is a hammer — manual intervention with eyes on the situation is usually better.
What happens to in-flight events
When a webhook auto-disables:
New events are not enqueued for it.
Events already in the queue continue through the consumer pipeline. They may still deliver before the disable propagates.
Replays of historical log rows still work — the disable affects new events, not manual replays.
Auto-disable vs license skip — different mechanisms
| Auto-disable | License skip |
|---|---|---|
Trigger | N consecutive HTTP failures on this webhook | Tenant-wide license state is |
Counts toward | A per-webhook counter | Nothing (excluded from stats) |
Resets on | First successful delivery | License renewal |
Visible state |
| Skipped neutral lozenge on log rows |
Recovery | Admin clicks Status toggle | Admin renews on Marketplace |
Both stop deliveries. Only auto-disable touches the webhook's state.
Common mistakes
Mistake | Effect | Fix |
|---|---|---|
Set threshold = 3 on a receiver with a 30-second-wide flaky window | Webhook auto-disables before failure pattern is meaningful | Raise to 10 or off. |
Re-enable an auto-disabled webhook without fixing the receiver | It auto-disables again on the next N failures | Diagnose first, then re-enable. See Re-enable an auto-disabled webhook. |
Expect Skipped (license) to count toward auto-disable | They never count | Renew the license; skipped rows aren't a webhook problem. |
Bulk-replay a hundred 5xx rows on a webhook near the threshold | Failed replays push it over | Be intentional with bulk-replay timing. |
Related
Tenant settings — the hub.
Retries and auto-disable — broader context, the manual-replay design.
Re-enable an auto-disabled webhook — the recovery playbook.
Webhook logs — where you see the failures that drive the counter.