Release 2.48.0-rc0#12959
Conversation
|
|
||
| ## 2.48.0-rc.0 / 2023-10-10 | ||
|
|
||
| * [CHANGE] Remote-write: respect Retry-After header on 5xx errors. #12677 |
There was a problem hiding this comment.
Wasn't sure about calling this a change or an enhancement, but it does change the current behavior of remote-write retries
There was a problem hiding this comment.
Could see it as a bugfix, even.
There was a problem hiding this comment.
Yeah, I just don't know how expected that behavior is. Especially as it's not a configuration option. But, I'm up to whatever you want.
Also on another note I see in that PR there was some discussion around documenting it, but I'm not sure if that that ever happened?
72853a7 to
0ac2007
Compare
|
did we update the dependencies? |
|
Was waiting on the go release to do that all in one |
Signed-off-by: Levi Harrison <git@leviharrison.dev>
0ac2007 to
827804c
Compare
bboreham
left a comment
There was a problem hiding this comment.
I had a couple of thoughts about making the text more widely understood.
Not critical.
|
Lgtm |
Signed-off-by: Levi Harrison <git@leviharrison.dev>
|
/prombench v2.47.2 |
|
Note that NewPossibleNonCounterInfo in #12152 makes rate queries inefficient, needs fixing before release. |
|
@zenador Looking at that PR again, that warning looks like something that we should be able to figure out statically when looking at the PromQL, so maybe it can be extracted from the actual function implementation completely? Then it also doesn't have to be evaluated (and the annotation added) for each resolution step of a range query. |
|
Addendum to doing it statically: It wouldn't work in the case where you don't equality-match on the metric name, but at least you could check it once after resolving series identities for selectors, not for each resolution step? |
|
Yes I agree, that particular warning and another one can be derived from the query itself, so we will move it out, but until then, we can remove it for now for a quick fix. However, other warnings do actually look at the data and so those will need to be applied every step. |
|
Benchmark tests are running for 3 days! If this is intended ignore this message otherwise you can cancel it by commenting: |
|
Benchmark tests are running for 5 days! If this is intended ignore this message otherwise you can cancel it by commenting: |
|
/prombench cancel |
|
Benchmark cancel is in progress. |




Waiting for 10/10 Go release