Memory leak in edge workers
We're seeing a memory leak in edge workers.
The memory graph in prod workers doesn't show the same steady, almost monotonic growth that edge has, though I do see one discrete jump.
I'm guessing that this the same issue we saw before; the request cache not being cleared in a background task
Steps to Reproduce
Reason for Variance
User Impact Summary
Closing this, since we don't seem to be affected by this anymore. It's odd that we'd see it one moment, but not another.
I don't see a memory leak on the workers over the past few days: https://rpm.newrelic.com/accounts/88178/servers/32935709/processes#sort_by=memory
: Talked with and briefly, and the consensus is that this doesn't appear to be an urgent issue (did not crop up over the weekend), and it can wait until you and your team return.
, I would actually love to look into this myself; but unfortunately I'm going to be on vacation next week. I was going to escalate this to , so can you and he work it out? If you think it can wait a couple weeks, as I said, I'd be really happy to look into it.
I think the thing we really need is a memory dump from devops when the memory usage grows a little bit over the next few days.
do you need support from Platform in investigating this?