feat(helm): add more specificity to affinity and topology spread constraints (#17404) - #19061
Conversation
f271766 to
b459b42
Compare
b459b42 to
00a8f8c
Compare
|
Hi @jkroepke - would you mind reviewing this please :) |
There was a problem hiding this comment.
| {{- tpl (.Values.adminApi.topologySpreadConstraints | toYaml) . | nindent 8 }} | |
| {{- tpl (. | toYaml) $ | nindent 8 }} |
this should work, too and we can keep the with. Right?
There was a problem hiding this comment.
@jkroepke - yeah thats much better than what I've done initially, thanks, I've updated the PR
5889619 to
ed3c4ef
Compare
ed3c4ef to
6d9beba
Compare
There was a problem hiding this comment.
sorry - I don't know if I should update this or whether you should do it.
There was a problem hiding this comment.
The CI automation updates the version, you don't need to do it.
There was a problem hiding this comment.
Thanks, I've reverted this change (and associated doc change).
6d9beba to
2c92c25
Compare
|
I'm not convinced that failing test is related to this PR, it passed last time. |
|
try update from main @TomHellier |
grafana#17404) This commit allows templating in the affinity and topology spread constraints, and updates the values to be more specific about the app.kubernetes.io/instance and app.kubernetes.io/name kubernetes recommended label. When deploying into a cluster with a low number of nodes, or a cluster with the app.kubernetes.io/component label already used extensively, you cannot place these pods. If we want them to be scheduled apart, we should be more specific with the label selector. This follows on from the draft PR grafana#17405
2c92c25 to
28627d5
Compare
|
@jkroepke I've rebased onto main |
…opology-spread-constraints Signed-off-by: Jay Clifford <45856600+Jayclifford345@users.noreply.github.com>
…opology-spread-constraints Signed-off-by: J Stickler <julie.stickler@grafana.com>
|
Great feature, but it seems the memcached template didn't include the necessary changes to use the syntax |
|
@eraac, it was missed unintentionally. I've disabled memcached in my instances of loki, and the helm charts are structured slightly differently so didn't come up when I grepped the source code. |
This commit allows templating in the affinity and topology spread constraints, and updates the values to be more specific about the
app.kubernetes.io/instanceandapp.kubernetes.io/namekubernetes recommended labels.When deploying into a cluster with a low number of nodes, or a cluster with the app.kubernetes.io/component label already used extensively, you cannot place these pods. If we want them to be scheduled apart, we should be more specific with the label selector.
This follows on from the draft PR #17405
What this PR does / why we need it:
Which issue(s) this PR fixes:
Fixes #17404
Special notes for your reviewer:
Checklist
CONTRIBUTING.mdguide (required)featPRs are unlikely to be accepted unless a case can be made for the feature actually being a bug fix to existing behavior.docs/sources/setup/upgrade/_index.mddeprecated-config.yamlanddeleted-config.yamlfiles respectively in thetools/deprecated-config-checkerdirectory. Example PR