Here is some log data that you may peruse for your...
# orm-helpm
Mike
08/15/2018, 7:33 PMHere is some log data that you may peruse for your enjoyment. It shows 2 crash-and-restart cycles and the accompanying log messages associated with said crashes.
n
nilan
08/15/2018, 7:48 PMPlease open a new discussion in the Forum and share as much information as possible about your setup and your query patterns 🙂
nilan
08/15/2018, 7:50 PMquick look at the logs suggests this is a CloudFront issue, what makes you think the issue is at the Prisma level?
nilan
08/15/2018, 7:50 PMDid you look into your server logs as well?
nilan
08/15/2018, 7:50 PMWhat does the monitoring for your DB say?
m
Mike
08/15/2018, 8:01 PMWe’re not using CloudFront, so that can’t be it.
DB monitoring shows that the DB is at normal memory, CPU, and connection levels.
Also, the log message from ECS when it restarts is:
service Prisma (port 60000) is unhealthy in target-group vibely-prisma-server-prisma due to (reason Request timed out).Mike
08/15/2018, 8:03 PMThe cloudfront calls there are calls to a server-side susbscription (webhook). The logs from those indicate that they’re failing because they try to call back to the Prisma server to perform updates or fetch more data and they receive
503 Service Unavailablen
nilan
08/16/2018, 7:18 AMThanks for creating https://github.com/prisma/prisma/issues/2926. Let's continue the discussion there 🙂