THIS IS A TEST INSTANCE. ALL YOUR CHANGES WILL BE LOST!!!!
...
In this example, AtMinIsr triggers when there is only 1 insync replica remaining, and tells us that 1 more failure will cause the partition to go completely offline!
Usage
This new AtMinIsr categorization can be extremely powerful in detecting broker failures (example 2 from above) and determining when action should be taken without being too noisy.
A potential usage of this new AtMinIsr category is:
- Set up an alert for AtMinIsr > 0 for a period of time
- If the alert is triggered, then assess the health of the cluster
- If there is broker failure which cannot be fixed quickly, then use partition-metric or --at-min-isr-partitions option of TopicCommand to quickly determine list of topics to repartition
AtMinIsr Values + Possible Explanations
...