Floofy System

We have found and addressed the issue that was causing Floofy.tech to periodically go offline over weekends. One of the tootctl cleanup commands we use for pruning data from our databases, specifically old/inactive/irrelevant accounts, was overwhelming our databases resources due to the sheer number of accounts in our database. We’re working with upstream Mastodon to find a good solution for everyone, but in the meantime we’ve imposed a limit to how many accounts will be scanned and pruned at a time.

We had a short outage as our database server went offline, and both admins are currently away at a convention. We’ve managed to reboot the database server and we’re catching up on federation events.

Root cause will be investigated Monday.

We have disabled a setting that compresses and converts images from our CDN to webp for most folks, which broke the [Pluraldawn](https://exa.y2k.diy/garden/pluraldawn/) userscript.

Generally speaking there should be little impact!

Beep boop, we've updated our instance to the latest glitch-soc commit. Enjoy!

We are placing a temporary domain Limit on https://mastodon.world due to the ongoing scampaign. We will revisit this Limit in the future.

:wyv:

There was a small disruption as our NFS server's disk filled up, which cause Redis to be unable to write its checkpoints. This caused it to crash. We are now recovering.

We are running some experiments on our infrastructure testing our current resource usage. This involves reducing the amount of resources we have allocated to the cluster to measure how much we *actually* need. We aren't expecting any performance impact, but are actively monitoring. If you see any weirdness, please let us know!

There are a large number of rety items in our Sidekiq queue, but federation to and from other instances appears to still be functional. From the logs, it looks like there are a handful of instances that are offline or unreachable. We will continue monitoring, but we have plenty of Sidekiq workers to work through the queue.

We will be upgrading our Mastodon instance shortly, and no downtime is expected - although with all upgrades, unexpected things can happen.

We have also changed how we manage our fork/minor Mastodon codebase customizations, which will now be published to https://github.com/Floofy-Mastodon/mastodon rather than as a collection of patchfiles. This will help us keep things up to date and modify behaviour down the road to make this instance truly ours :3

We'll be pushing out a minor refactoring of our Kubernetes state git repository soon. No impact is *expected*, but it's slightly higher risk and we will be monitoring very closely.

Mastodon v4.5.0 has been released! We've merged the changes into our fork and we will be updating our instance later today. https://github.com/mastodon/mastodon/releases/tag/v4.5.0

No impact is expected :3

We will be upgrading our instance shortly! You can check out some of the changes in the git log https://github.com/Floofy-Mastodon/mastodon/commits/main/.