Scheduled jobs

We’ve several different scheduled jobs across our infrastructure:

• Cronjobs on the Linux servers

• systemd timers on the Linux servers

• Scheduled GitLab CI pipelines

• Application tasks (e.g. via Celery)

Schedules

Important

To ensure a stable environment, it’s important that all the jobs are properly synced in time. Whenever setting up a new job, ensure you’re following the schedules below.

OS schedules

Time	Weekday	Action	Source
*:00 (+ random 15min)	everyday	Run borg backups.	borg role
04:30 (+ random 15min)	everyday	Download apt upgrades. After CI jobs, before upgrades.	apt_unattended_upgrades role
05:30 (+ random 5min)	everyday	Install apt upgrades. After CI jobs & downloads, before morning.	apt_unattended_upgrades role
05:50 / 05:55	Sundays	Reboot systems when required. After upgrades, before morning.	reboot role

CI & Docker schedules

Time	Weekday	Action	Source
/4:00	everyday	Run GitLab Guard. Doesn’t affect backup much.	GitLab Guard pipeline
/4:00	everyday	Run Renovate. Doesn’t affect backup much.	Renovate pipeline
/4:00	everyday	Run Docker Registry garbage collection. Doesn’t affect backup much.	Docker registry pipeline
03:30	everyday	Build Docker Base images (stage 1). After backup is done.	Docker image pipelines
03:45	everyday	Build Docker Base images (stage 2). After stage 1 is done.	Docker image pipelines
04:00	everyday	Run library tests. After new base images.	Python library pipelines
04:30	everyday	Run Docker system garbage collection. After backup, far after new base images.	docker-gc timer docker-gc service

Docker service schedules

Time	Weekday	Action	Source
hourly (+ random 5min)	everyday	Database backups.	Postgres Docker image entrypoint MariaDB Docker image entrypoint
hourly	everyday	Run LDAP backup.	LDAP Docker image entrypoint
04:45	everyday	Run GitLab backup. After CI jobs, before next backup.	GitLab Docker image entrypoint