This report covers the San Francisco Bay regional Mastodon instance sfba.social, with general statistics, financial details including income (donations) and expenses (hosting costs), moderation efforts, and changes made and being considered.

General Statistics

On January 31st, 2023 we tallied roughly 37,500 friends on sfba.social. This is an increase of about 750 new users during the month, quite a lot fewer than the 5000 new friends we gained in December.

From Jan 1st-31st, the moderation team received 95 reports, about 25% of December’s total.

The Mastodon back end counted roughly 529,000 interactions, about 50% of what we saw in December, 2022.

Financial Details

We are very fortunate in that financial contributions are easily covering the monthly hosting costs. We are continuing to build reserves to sustain the instance long-term. 

Our expenses are mainly monthly hosting costs with Linode and a steadily decreasing cost with AWS:

Our very simple balance sheet:

With our projected burn rate, the running costs of the instance are covered for over five more months now. We are targeting 6-months of funding reserve.

Donations are gratefully accepted at: opencollective.com/sfba-mastodon

Moderation

Moderators work to make sure that the local feed posts follow sfba.social rules. Our moderation efforts rely on users promptly reporting posts that go against our server rules, which helps us quickly deal with issues as they come up. Users are encouraged to actively curate their feeds using Filters and the Mute and Block tools. 

The Admin/Mod team handled 95 reports in the month of January. This is a decrease over December, where the moderation team resolved 364 reports. 

Before November, 2022 sfba.social averaged under 1 report/100 users. Then came the big influx and the number of reports fluctuated:

• Nov. jumped up to 2.14 reports/100 users (31800 users)
• Dec. moved back to 0.9 reports/100 users (36750 users)
• Jan. is down to 0.25 reports/100 users (37500 users)

The vast majority of reports are resolved within 24 hrs.

At the end of the month, there were 0 open local reports and 0 reports on remote instances, as well as 1 pending appeal. 

We actively address bad actors on the federated network by defederating instances with poor moderation standards or a history of lax enforcement. We monitor the #fediblock hashtag to identify servers that may warrant investigation on our part, before they become an issue for our users. 

A list of defederated servers on our about page (under “Moderated servers”). Servers marked as “limited” cannot communicate with our server by default, but users can follow and communicate with specific accounts when they know the full account address. Servers marked as “suspended” cannot communicate with sfba.social, with no exceptions.

Team

No changes: the Admin/Mod team is stable and the work load is manageable. 

Events

The team has been working on an in-person meet-up for the whole community. Club Fugazi (@ClubFugaziSF) stepped up with a suggestion and now we have organized an in-person event for our members to meet the admins and have a local entertainment experience at a reduced price! (A clear benefit of a geographically local instance is that we are (mostly) co-located, enabling us to coordinate local events.)

We are interested in continuing to organize local meet-ups and creating more opportunities for our members to meet each other! We want to hear from you: what kind of in-person events would you like to see?

Improvements

Implemented

• Increased reliability: besides the fact that we’re using just one region of one cloud provider, our architecture had just one last remaining single point of failure: the database server. We considered using a managed cluster, however, it seemed to be cost prohibitive. Instead we implemented a database cluster on top of 3 regular VMs. One of the 3 is the primary, the two other ones are hot replicas. If the database on the primary were to fail, the two secondaries would agree on a new leader and take over, this would happen within a few seconds. If not only the database application but the entire node were to fail, one of the secondaries would take over the IP address of the primary and continue serving, this would take about 30 seconds. For those interested in the details, we followed this guide pretty closely: https://www.stepstoperform.com/2022/11/pgsql14-ha-cluster-ubuntu20.html
• Matrix for messaging: we started collecting experience with the configuration, operation, and administration of matrix. We’re considering opening it up to the sfba community, but need to overcome some technical hurdles first, notably single sign on. This technology would provide a private direct-messaging service for our users, without requiring a new account. 

In Progress

• Adding an on-boarding guide to the information hub to help new users get going quickly. 
• Increase of the character limit on posts from 500 to 1024. 
• Adding additional language to the Code of Conduct to provide ways for local businesses to engage without overwhelming the timeline and allowing users to opt-out. (Would be announced via @announcements account).

Under Consideration

• Switching code base from vanilla Mastodon to Hometown. If we decide that these things are doable and maintainable, we will involve the SFBA community in the decision making process, of course. 

We look forward to the next month/year– If you would like to contribute, please donate to opencollective.com/sfba-mastodon and we always welcome comments!

Your Admin & Moderation team: 
@seb @cd24 @moritz @ingurido @jeffkibuule @neuralgraffiti