27/01/2025 18:19 GMT Some VPS rebooted
Our infrastructure deployed the new version of our hypervisor in the host machine in order to prepare for the next upgrade. However, instead of downloading this new version from our binary cache, some host machines started to compile it from source. This compilation triggered OOM kill of the VM process (indeed the VM process is configured to use almost all RAM from the host machine). Of course, slapos detected this problem right away and restarted the VM process so the downtime was minimal.
The reboot happened between 18:19 GMT and 18:55 GMT.
We are sorry for this inconvenience.
We already fixed the configuration of all host machines to correctly force the download from binary cache.