[tor-project] minutes from the sysadmin meeting

Hello everyone,

Here’s your again regular dose of sysadmin meeting goodies.

Roll call: who’s there and emergencies

anarcat, gaba, lavamind. kez AFK.

Dashboard cleanup

Normal per-user check-in:

General dashboards:

Delegated web dashboard review to gaba/anarcat sync on Thursday. We
noticed we don’t have the sponsor work in the roadmap page, we’ll try to
fix this shortly.

Vacations planning

Discussed the impact of the unlimited PTO policy which,
counter-intuitively, led some team member to schedule less vacation
time. There are concerns that the overlap between anarcat and lavamind
during the third week of july could lead to service degradation or
delays in other deliverables. Both lavamind and anarcat have only
scheduled “PTO” (as opposed to “AFK”) time, so will be available if
problems come up.

There should probably be a discussion surrounding how emergencies and
availabilities are managed, because right now it falls on individuals to
manage this pressure and it can lead to people taking up more load than
they can tolerate.

Metrics of the month

  • hosts in Puppet: 86, LDAP: 85, Prometheus exporters: 156
  • number of Apache servers monitored: 35, hits per second: 652
  • number of self-hosted nameservers: 6, mail servers: 8
  • pending upgrades: 111, reboots: 2
  • average load: 0.74, memory available: 3.39 TiB/4.45 TiB, running processes: 588
  • disk free/total: 36.98 TiB/110.79 TiB
  • bytes sent: 316.32 MB/s, received: 206.46 MB/s
  • planned bullseye upgrades completion date: 2023-02-11 (!)
  • GitLab tickets: 193 tickets including…
    • open: 0
    • icebox: 147
    • backlog: 22
    • next: 9
    • doing: 10
    • needs review: 1
    • needs information: 4
    • (closed: 3164)

Upgrade prediction graph lives at

The completion date is still incorrect, but at least it moved ahead in
time (but is still passed).