In this chapter, we have covered the maintenance concepts of Storm in terms of adding new nodes, rebalancing, and killing topologies. We have understood and tweaked internals such as numtasks and parallelism in combination with to numworkers and network latency. You learned to locate and decipher logs of Storm components. You also understood the metrics of the Storm UI and their implications on topology performance.

In the next chapter, we will discuss advanced concepts of Storm, including micro-batching and Trident APIs.