All services are online_

We have increased the capacity of the affected database proxy, and response times are returning to normal in all regions. We're monitoring the recovery and will post an update once it is complete.

API requests in the Frankfurt region and the Console slowed down, and some timed out. Appwrite Functions and Sites in all regions were also slower, and some requests failed. A sudden rise in traffic exhausted capacity on an internal database proxy that these requests depend on.

This incident was recorded retrospectively. Realtime subscribers in the Frankfurt region stopped receiving events. Realtime connections and subscriptions still succeeded, and API requests, including the writes that produce events, completed normally.

We're investigating intermittent slowness in Storage in the San Francisco region. Some file operations take several seconds longer than usual to complete. Uploads, downloads and file previews continue to succeed. We're working with our infrastructure provider and will share updates here.

While we prepare the fix, you can work around the issue by switching your site from the ephemeral key to a standard API key with the scopes your site needs.

We have identified the cause. A proactive security hardening change we made introduced a regression that stops ephemeral keys from working. We are preparing a fix and will post an update when it is rolling out.

We're investigating degraded performance for Appwrite Sites in all regions. Ephemeral keys are not working, so some sites cannot complete requests that depend on them. We will post an update as soon as we know more.