Previous incidents

Meta have confirmed the update is resolved

From - https://metastatus.com/whatsapp-business-api
"We have recovered from an earlier outage impacting Cloud API, and services have now been restored. We apologize for any inconvenience that this may have caused."

We have spoken to OVH who confirm the routing was adjusted by the upstream carrier not by them.

The API has returned to normal.

Any messages would have been queued our side and retried during the disruption.

Update from Postmark: We've deployed additional fixes and latency/errors have returned to normal performance. Engineering will continue to monitor into the evening for consistent results.

Postmark have confirmed issues with their API which they now believe it to be mitigated.

You can find more information at https://status.postmarkapp.com/notices/iejilmig3r9ldil5-increased-api-latency-timeouts

Around 5% of workflows which included a specific timed delay may have been delayed for up to 24 hours longer than the time specified. This issue was caused by a misconfiguration.

No workflows were missed, and all affected workflows would have continued as planned, albeit with a larger than planned delay.

The situation has continued to remain stable.

During the disruption period, 1.04% of requests resulted in a timeout due to the network issues according to 3rd party monitoring.

Cloudflare has posted a good summary of the incident from their perspective - https://blog.cloudflare.com/cloudflare-perspective-of-the-october-30-2024-ovhcloud-outage/

This has been confirmed as resolved.

Between 13:14 UTC and 13:19 UTC there was a brief network connectivity issue between Azure LHR, Azure JNB and OVH LON where connections where stalled or timed out.

For any customer whose voice is hosted in one of those Azure regions, but data is held in OVH London - this may have resulted in disruption to recordings.

The issue was automatically resolved when the network connectivity resumed.

This issue did not affect any other region than the ones listed.

This issue reoccured briefly between 11:41 and 11:49. Circuit breaker limits further increased as well as now being fully automated.