setup a cache frontend for the blog
design docs in https://help.torproject.org/tsa/howto/cache/
launch checklist:
- alternatives listing and comparison (done)
- deploy a test virtual machine by hand, say
cache-01.tpo
(done) - benchmark the different alternatives (done, ATS and nginx comparable)
- setup secondary node with Puppet, say
cache-02.tpo
(done) - validation benchmark against both nodes (done)
- lower DNS to 10 minutes wait an hour (done)
- open firewall (done)
- lower DNS to 3 minutes (done, around 2019-11-05 16:00:00)
- point DNS to caches (done)
- raise DNS back to 1h if all goes well. (done!)
Post launch tasks:
- update documentation to fill in information from template (done)
- hit ratio stats in to Prometheus, separate ticket? (done, although missing syslog buffer, see #32461 (moved))
- convert existing varnish setups into Nginx (probably requires Puppet refactoring, see #32462 (moved))
Disaster recovery:
- flip DNS back to backend