Decide what service failures count as an emergency for anti-censorship team
In the context of upcoming vacation time, Gaba asked us a great question: What do we consider an emergency in the anti-censorship team?
For comparison, the sysadmin team tried to figure out what is "top priority" and ended up with an answer like "www website; donate website; email" where if those things catch fire they'll drop everything to address them, whereas if anything else catches fire it'll have to wait until a TPA person gets to it. More details here:
https://gitlab.torproject.org/tpo/tpa/team/-/wikis/policy/tpa-rfc-2-support#code-red
So it would be a good exercise for the anti-censorship team to (a) enumerate the important services and features and other things that we support -- "things that can fail and need attention" -- and then (b) sort them into a few buckets by how urgently we want to respond if they break.
Goals include:
-
Have clear answers on which failures to escalate and which ones can wait. (For example, do you want us to interrupt your vacation if Russia does another DPI-based Snowflake block? How about if a country with few Tor users does one? My guess is 'yes' and 'no' respectively.)
-
Help prioritize writing / updating survival guides for the top items.
-
Help prioritize fixing developer / access / knowledge bottlenecks, e.g. moving everything from "one person can do it" to "two people can do it."
-
Letting us compare our list with lists from other teams to discover surprises. For example, if we decide that moat is a critical service for us, but the sysadmin team has decided that availability of the moat server isn't on their critical list, then we can start those conversations.
We don't need to rush to get answers for this particular 2021 holiday, since we can continue to work by using the intuition of whoever is around and notices a problem. That is, consider this ticket "important" but not "urgent".