Being a good admin

Status: placeholder. This guide will be written before launch.

The closing guide in this section. Not a manifesto and not a list of platitudes — a short, concrete piece on the practices that make calendar owners want to stay on your instance and other admins want to federate with you. Each item anchored to a moment of decision, per the rest of the guides.

Planned scope

• Answer mail. This is most of the job. The admin who responds to email within a day, even with "I'll get back to you," is the admin people trust
• Be honest about outages, in writing. A two-sentence "the instance was down for 40 minutes this morning because of X" is worth more than silence and more than spin
• Post-mortems when something breaks, even short ones. Public radio publishes its numbers; you publish what went wrong. Same principle
• Don't surprise people with policy changes. Announce, give a window, then act. The thirty-day notice on a defederation is almost always cheaper than the trust cost of doing it the same day
• The relationship with other admins is part of the job, not extra. The admins on the instances you federate with are colleagues, and the time you spend treating them that way pays back in the cases where you need their cooperation
• Sustainability for the admin. Running a community instance can be a long-term thing or a short-term thing; both are fine, but the difference is whether you've planned the handoff. The admin who burns out without a successor leaves a vacancy that the community can't easily fill
• The "I'm not sure I'm doing this right" question. The honest answer is that nobody starts out doing this right. The signal is whether you adjust when you find out you weren't