We've upgraded the postgresql database to a $240/mo VM. This has solved today's latency. We will continue to monitor the instance but believe we're in good shape to handle tomorrow's traffic.

Latency remains around two hours however good news: queues are starting to recover due to some actions taken.

Actions taken:
- pg_repack on the database (took 45 minutes!)
- pulled fresh code from the git repository and recompiled both the web and job processing servers.

We are experiencing latency of up to two hours. You may also find that at times third-party apps may not be able to connect to the instance. This is expected as we try to get queues back to a manageable level.

We recognize that this is frustrating and are working from every angle we can to get this under control.

We're aware that mstdn.ca is running slow. We thank you for your patience.

Nous savons que mstdn.ca fonctionne au ralenti. Merci pour votre patience.