Is Materialize down?

Between 21:12 UTC and 22:18 UTC on September 3, AWS experienced elevated error rates in its Security Token Service (STS) in the us-west-2 region, which AWS reported on its Health Dashboard. Materialize environments in that region use STS to obtain credentials for object storage access, and requests for those credentials failed or timed out during the event. Some environments in the region experienced degraded performance during this window. Query results remained correct throughout, and no data was lost. The upstream issue was resolved by AWS at 22:18 UTC and all affected environments recovered.

We are seeing issues across the platform potentially caused by an outage of an upstream vendor. We are actively investigating and looking into mitigations.

We’ve determined that this issue is related to an outage of our authentication provider and is limited to pgwire authentication. Sources and sinks continue to run as expected, and we do not expect this issue to impact data freshness. We are continuing to monitor the situation and work toward a resolution.

We’re seeing pgwire authentication recover and are continuing to monitor the situation to ensure service remains stable.

Upstream authentication vendor has fully resolved their incident. All impacted systems have recovered.