Is log cleaner logging anything on AWS MSK?
I use AWS MSK cluster with brokers logging turned on to CloudWatch. I can see brokers logs. We have some topics with cleanup.policy=compact and some with cleanup.policy=delete. The system is running on the new cluster for about 2 weeks now.
From my research (e.g. https://zendesk.engineering/an-investigation-into-kafka-log-compaction-5e520f4291f0) I see that kafka should run log cleaner (obviously) and there should be some traces in logs of this activity. However in my CloudWatch log group I cannot find a word "cleaner" or "cleaned" and I cannot find any trace of log cleaner running.
Is it something wrong with my logging settings? Or is log cleaner not running at all (this is suspicious, because we have a lot of messages eligible for cleanup but still not cleaned)?
Hello there, in order to verify from the logs whether files related to partitions are getting deleted or not you can look for the something similar as below
Deleted log LOG_DIR/<topic>/00000000022134078627.log.deleted
Deleted time index LOG_DIR/<topic>/00000000022134078627.timeindex.deleted
When retention kicks in you will be noticing deletion of log and time index files in the broker logs. That indicates retention is successfully triggered and deleted old log segments.
Relevant questions
Scaling a cluster
asked 3 years agoIs log cleaner logging anything on AWS MSK?
asked 2 months agoAmazon MSK Authentication and Authorization
asked 6 months agoClarifying MSK data transfer pricing within region or AZ
Accepted Answerasked 3 years agoStorage Capacity on MSK maxed out
asked a month agoUnable to use IAM permissions to access MSK Brokers
asked a year agoLogging with CloudWatch vs. ElasticSearch/Kibana
Accepted Answerasked 5 years agoI want consolidated application logs running on AWS ECS with microservices architecture
asked a month agoAutomatically update brokerstring in MSK
Accepted AnswerBrokers Connection string doesn't show creating cIuster with IAM Access
Accepted Answerasked a year ago
Thank you! Yes I have noticed such logs, but still, the disk usage is constantly growing, while we have about 99% of records eligible for deletion. This records are in __consumer_offsets topic where we commit last message every 100 millis even if there are no new messages for hours.