2 months ago
My CPU and Memory has started going up this month despite my traffic being the same. And there has been no change in the code (it's the same docker image from Hasura as before). Maybe Hasura has a memory leak, but it's a battle-tested service that performed great all last month. I could also try their latest patch update docker image, but I wanted to show this to Railway first since it doesn't look like a Hasura issue to me.
Attachments
Pinned Solution
2 months ago
Yep, probably a memory leak or some accumulated resource like unclosed connections, before doing a cron to redeploy it once a week I suggest you update the source image to see if it fixes this leak and check if there aren't any warnings in the logs pointing to a possible cause.
4 Replies
2 months ago
I redeployed, and it's back down to normal levels. So does that point to it being a Hasura related issue? I could setup an automation to re-deploy about once a week.
joshuadutton
I redeployed, and it's back down to normal levels. So does that point to it being a Hasura related issue? I could setup an automation to re-deploy about once a week.
2 months ago
Yep, probably a memory leak or some accumulated resource like unclosed connections, before doing a cron to redeploy it once a week I suggest you update the source image to see if it fixes this leak and check if there aren't any warnings in the logs pointing to a possible cause.
Status changed to Solved brody • 2 months ago