How do graphQL engines like Hasura know to trigger jobs once even when they're scaled to multiple instances?

Question

I deployed a few instances of the same Hasura server unto AWS ECS, I added a CRON job that I wanted to test out, and apparently the job is run only once (which is the ideal situation).

I was worried I'd have to figure out how to make sure jobs are run only once when I scale an instance horizontally, I'm not complaining, I would just like to know how that works, because I couldn't find it in the documentation.

score 1 · Accepted Answer · answered Feb 05 '22 at 14:07

I can't find it in the documentation either, but Hasura appears to track scheduled cron jobs in the database, and I assume it is using a database locking mechanism to prevent a single job from running on more than one server at a time. Just like other similar job scheduling tools.

How do graphQL engines like Hasura know to trigger jobs once even when they're scaled to multiple instances?

1 Answers1