[jira] [Created] (KAFKA-13384) FailedPartitionsCount metric is not updated if a partition log file was corrupted

["Alexander (Jira)" <ji...@apache.org>]

Alexander created KAFKA-13384:
---------------------------------

             Summary: FailedPartitionsCount metric is not updated if a partition log file was corrupted
                 Key: KAFKA-13384
                 URL: https://issues.apache.org/jira/browse/KAFKA-13384
             Project: Kafka
          Issue Type: Bug
    Affects Versions: 2.6.2
         Environment: OS:
NAME="Amazon Linux AMI"
VERSION="2018.03"
ID="amzn"
ID_LIKE="rhel fedora"
VERSION_ID="2018.03"
PRETTY_NAME="Amazon Linux AMI 2018.03"
CPE_NAME="cpe:/o:amazon:linux:2018.03:ga"
HOME_URL="http://aws.amazon.com/amazon-linux-ami/"

Kafka version:
2.6.2 (Commit:da65af02e5856e34)
            Reporter: Alexander
         Attachments: Screenshot 2021-10-19 at 15.28.33.png

We found a misbehavior on our Kafka cluster (version: 2.6.2 (Commit:da65af02e5856e34)), `FailedPartitionsCount` metric is not updated if a partition log file was corrupted

Steps to reproduce the problem:
1. corrupt a partition log file
2. restart Kafka process

After that, you will get a correct log which tells that Kafka marked corrupted partitions as failed

```
2021-10-19T14:49:31+02:00 [2021-10-19 12:49:30,924] WARN [ReplicaFetcher replicaId=11, leaderId=10, fetcherId=0] Partition test_topic-1 marked as failed
 (kafka.server.ReplicaFetcherThread)
```

But the value of `FailedPartitionsCount` metric will be 0 (see attached screenshot)



--
This message was sent by Atlassian Jira
(v8.3.4#803005)