Before you reach for a job queue, a scheduler service, or a cloud function that costs money to do nothing 99% of the time — ask whether a cron line would've done it. Usually it would have.

# m h dom mon dow  command
0 4 * * *  /opt/bin/rotate-logs && /opt/bin/sync-offsite

The one upgrade I always make: never let cron email me silently into the void. I wrap jobs so a failure actually pings me, because a backup job that's been quietly dying for three weeks is worse than no backup job at all — it's a backup job and a false sense of security.

cron has been doing this faithfully since before I was born. It is not the bottleneck in your architecture. It is the most reliable thing you own.