-
Notifications
You must be signed in to change notification settings - Fork 548
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Jsonnet / Helm: relax the hash ring heartbeat period and timeout for distributor, ingester, store-gateway and compactor #6860
Conversation
This comment was marked as outdated.
This comment was marked as outdated.
53d2ce2
to
64e9bef
Compare
This comment was marked as outdated.
This comment was marked as outdated.
5f09fa3
to
da29aad
Compare
…ngester, store-gateway and compactor. These are values which help reduce the pressure on a KV store or reduce the CPU spent by memberlist in passing messages. The tradeoff is that abrupt shutdowns/crashes of components will take longer to detect by peers. We've been running with these values at Grafana Labs for some time and haven't seen problems. Signed-off-by: Dimitar Dimitrov <[email protected]> Signed-off-by: Marco Pracucci <[email protected]>
cfb91ca
to
a677f2f
Compare
Signed-off-by: Marco Pracucci <[email protected]>
Signed-off-by: Marco Pracucci <[email protected]>
Signed-off-by: Marco Pracucci <[email protected]>
Signed-off-by: Marco Pracucci <[email protected]>
Signed-off-by: Marco Pracucci <[email protected]>
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
thanks for driving this to completion @pracucci. In principle LGTM. Reviewing this code raises questions that I didn't ask myself before 😅
operations/compare-helm-with-jsonnet/components/config/kustomization.yaml
Show resolved
Hide resolved
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM (even though I opened this PR)
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I took ownership of this PR from Dimitar. I did the changes on Dimitar's PR, so I've asked Dimitar to review it. Dimitar approved, but can't officially approve its own PR, so I'm approving it (but it's not a self-approval, it's Dimitar approving it in practice).
In this PR we're upstreaming a config change we've done long time ago at Grafana Labs, relaxing the hash ring heartbeat period and timeout for distributor, ingester, store-gateway and compactor.
These are values which help reduce the pressure on a KV store or reduce the CPU spent by memberlist in passing messages.The tradeoff is that abrupt shutdowns/crashes of components will take longer to detect by peers. We've been running with these values at Grafana Labs for some time and haven't seen problems.
Checklist
CHANGELOG.md
updated - the order of entries should be[CHANGE]
,[FEATURE]
,[ENHANCEMENT]
,[BUGFIX]
.about-versioning.md
updated with experimental features.