upgrade thanos that include skip bad blocks during compaction #4453
Conversation
This isn't a good way to express the changes - "since" will change every day. Also put a short description in the CHANGELOG, mentioning the new metric. |
huyan0
force-pushed
the
master
branch
3 times, most recently
from
6e3f5cf to
f8119ca
Compare
huyan0
changed the title
CHANGELOG.md
Outdated
| @@ -1,7 +1,7 @@ | |||
| # Changelog | |||
|
|
|||
| ## master / unreleased | |||
|
|
|||
| * [CHANGE]: Compactor: Introduce a new metric to monitor blocks marked for no compact during compaction: cortex_compactor_blocks_marked_for_no_compact_total | |||
There was a problem hiding this comment.
Would be nice if we can briefly explain why when would cortex_compactor_blocks_marked_for_no_compact_total go up (i.e why would block get marked for no compact). This will give the Cortex admin, as well as customers reading the change log more context on the background of the metrics.
|
|
||
| // TSDB syncer metrics | ||
| syncerMetrics *syncerMetrics | ||
| } | ||
|
|
||
| type noOpsCounter struct { |
There was a problem hiding this comment.
based on sync offline, I have created this noOpsCounter so that there's no metric when feature is disabled. I don't think this is elegant and I'm looking at some examples. Would appreciate some advice.
There was a problem hiding this comment.
@bboreham @pracucci @pstibrany it would be nice to have your thoughts regarding this as well - should we not emit the metric at all when "skip bad block" feature is disabled.
There was a problem hiding this comment.
You can just pass prometheus.NewCounter(prometheus.CounterOpts{}) - if it isn't registered it doesn't go anywhere.
Signed-off-by: Yang Hu <[email protected]>
|
|
||
| // TSDB syncer metrics | ||
| syncerMetrics *syncerMetrics | ||
| } | ||
|
|
||
| type noOpsCounter struct { |
There was a problem hiding this comment.
You can just pass prometheus.NewCounter(prometheus.CounterOpts{}) - if it isn't registered it doesn't go anywhere.
|
when this PR is going to be release? |
|
This issue has been automatically marked as stale because it has not had any activity in the past 60 days. It will be closed in 15 days if no further activity occurs. Thank you for your contributions. |
|
We also ran into the issue described by #3569. Is there something I can do to help with this PR? |
|
This issue has been automatically marked as stale because it has not had any activity in the past 60 days. It will be closed in 15 days if no further activity occurs. Thank you for your contributions. |
|
I will close this as it was implemented on #4707 |
Signed-off-by: Yang Hu [email protected]
What this PR does:
This PR is a follow up of #4423. Following the recommendation, I created this PR to update Thanos dependency before supporting skipping bad blocks with bucket index enabled and creating a feature flag.
Update Thanos version to include 4566 and 4469 to skip blocks with out-of-order chunks. The skip is currently disabled and will be enabled in a follow-up PR
Which issue(s) this PR fixes:
Step towards fixes #3569
Checklist
CHANGELOG.mdupdated - the order of entries should be[CHANGE],[FEATURE],[ENHANCEMENT],[BUGFIX]