Add metrics to measure: * histogram of event batch sizes * db write latency normalized by batch size
Add metrics to measure: