a year ago
Hi! When deploying my projects, I get a temporary 502 error that lasts for about 20sec. This disrupts services for my customers.
I have a single instance running Django + Postgres
I have properly setup the health endpoint, but this has not really helped - maybe it has shortened it by 10 sec.
I am using Cloudflare, so customers see the Cloudflare 502 page.
Is there a way to minimise or remove this 502 error state on a single instance? I am coming from Heroku, and I never saw this issue there.
Thank you!
3 Replies
a year ago
Two deployments can not be attached to the same volume, this is done to mitigate data corruption, thus however long the new deployment takes to come online would be the amount of downtime you see, since the previous deployment is stopped before the new deployment is started.
Have your application start faster.
Instead of a volume, use external storage, like minio.
a year ago
Ah, thanks! I would never have guessed it was the volume. Yet another reason to not use volumes in production - I am already planning a move to external storage.
Status changed to Solved brody • over 1 year ago