We have been seeing a number of timeouts for offset commits: `Offset commit failed on partition mytopic-12 at offset 16578269: The request timed out.`.
These failed requests are sporadic and not causing a problem "yet", but we would rather like to understand what could the causes of something like this be and what sort of tuning might resolve this or reduce the number of these timeouts. If I am reading the code right, this error is coming from one of the brokers (most likely during replication). We do see that other offsets are committed without an issue in that request. Our cluster consists of 3 brokers and has a `min.insync.replicas` value of 2. This only happens less than 10x per day and appears that on retry works just fine so it's not a critical issue. Thanks in advance, Matt
