the commit is a part of the bigger work to create multi-timeline consumer pool for subscriptions statistics. that's needed because in order to get statistics, the consumer has to be created for every partition for every request and read actual events. with the bigger load the amount of consumers created per second can reach very big numbers, which affects Kafka performance negatively and creates issues on another endpoints for example increased publishing latency. in order to avoid that without changing the architecture of the statistics endpoint, the pool of consumers can be created and consumers can be safely reused. and here comes the current commit, which enables us to reuse the multi-timeline consumer utilising reassign method otherwise consumer is thrown away after every assigning of new topic partition.
One-line summary
reassign topics in multitimeline consumer
the commit is a part of the bigger work to create multi-timeline consumer pool for subscriptions statistics. that's needed because in order to get statistics, the consumer has to be created for every partition for every request and read actual events. with the bigger load the amount of consumers created per second can reach very big numbers, which affects Kafka performance negatively and creates issues on another endpoints for example increased publishing latency. in order to avoid that without changing the architecture of the statistics endpoint, the pool of consumers can be created and consumers can be safely reused. and here comes the current commit, which enables us to reuse the multi-timeline consumer utilising reassign method otherwise consumer is thrown away after every assigning of new topic partition.