Thanks everyone for the feedback and votes. I have three +1s (David, Colin,
Jose).
Closing this vote now.
On Thu, Aug 4, 2022 at 2:09 PM José Armando García Sancio
wrote:
> Thanks for the improvement. LGTM. +1 (binding).
>
> --
> -José
>
- Niket
Thanks for the improvement. LGTM. +1 (binding).
--
-José
Hey Jose,
> How about the inactive controller? Are inactive controllers going to
update this metric when they encounter an error when replaying a
record?
Yes, this metric will be reported for both active and inactive controllers.
The Inactive controllers will update this metric when they encount
Thanks for the KIP Niket.
> kafka.controller:type=KafkaController,name=MetadataErrorCountReports the
> number of times this controller node has renounced leadership of the metadata
> quorum owing to an error encountered during event processing
How about the inactive controller? Are inactive con
Thanks, Niket. +1 (binding)
Colin
On Wed, Aug 3, 2022, at 15:24, Niket Goel wrote:
> Thanks for the explanation Colin.
>
>> ForceRenounceCount =>
>> kafka.controller:type=KafkaController,name=MetadataErrorCount
>> publisher-error-count => metadata-load-error-count
>> listener-batch-load-error-co
Thanks for the explanation Colin.
> ForceRenounceCount =>
> kafka.controller:type=KafkaController,name=MetadataErrorCount
> publisher-error-count => metadata-load-error-count
> listener-batch-load-error-count => metadata-apply-error-count
Yeah, this makes sense. I have made the changes in naming
I think there are a few different cases here:
1a. We hit an ApiException PREPARING metadata records on the active controller.
This is normal and expected. For example, someone tried to create a topic that
already exists. We translate the ApiException to an ApiError and return the
appropriate er
Thanks for taking the time to go over the KIP Colin.
While I agree with both your points about error handling, I think this KIP
focuses on just exposing these errors via the proposed metrics and does not
alter the error handling behavior on either the brokers or the controllers. The
metrics (as
Hi Niket,
Thanks for the KIP -- much appreciated! The new metrics look very useful.
I agree with the proposed error handling for errors on standby controllers and
brokers. For active controllers, I think we should establish two points:
1. the active controller replays metadata before submitting
Thanks, Niket.
+1 binding from me
-David
On Mon, Aug 1, 2022 at 8:15 PM Niket Goel wrote:
>
> Hi all,
>
> I would like to start a vote on KIP-859 which adds some new metrics to KRaft
> to allow for better visibility into log processing errors.
>
> KIP
> —ttps://cwiki.apache.org/confluence/dis
10 matches
Mail list logo