[INT-200] Check database connection on startup and on /healthz endpoint
#167
Conversation
…oint for monitoring
archessmn
force-pushed
the
int-200-healthz-should-check-database-status
branch
from
03bb374 to
18be8a3
Compare
|
Deployed a preview of this PR to https://pr-167-internal.dev.ystv.co.uk |
1 similar comment
|
Deployed a preview of this PR to https://pr-167-internal.dev.ystv.co.uk |
| console.error( | ||
| `Database connection attempt ${connectionAttempts} failed, retrying...`, | ||
| ); | ||
| await sleep(5000); |
There was a problem hiding this comment.
Should we bail immediately and not do the whole retry loop / is this delaying the inevitable? Would also simplify this code substantially.
There was a problem hiding this comment.
My reasoning was just to give it a bit of time to get started so it wouldn't just fail instantly and keep getting redeployed if it had an issue on the first connection
There was a problem hiding this comment.
Possibly? But in my head that'd only be the case if the database server is the one that's taking a while to start, which it should almost never be in production (the Prisma connection should block if it isn't ready yet), so we'd want to fail fast if something is up with the DB.
There was a problem hiding this comment.
It's within 15 seconds, it's not going to cause a massive issue I don't think
Do database connection monitoring slightly better