We're currently improving our SSL infrastructure, automating Let's encrypt renewals on Dwellers.
This automatic certificate renewal means we need to restart web server. The script to restart the webserver contained a logic error, leading to spawn a lot of processes (a fork bomb) during a test. As such, Dwellers experienced some issues.
19:36 UTC — Perturbations on Dwellers, some services are lagging or not reachable 19:44 UTC — Fork bomb is defused. CPU and schudeling back to normal. 19:44 Retour à la normale 19:46 UTC — MySQL container has been restarted. All services excepted Etherpad are back to normal
21:54 UTC — Production environment tests are launched. We detect Etherpad container was down too, restarted. Production environment tests pass again.