#6613 Docs: New section for `Replicating Data > Troubleshooting > Performance Benchmarking` Issue created by aaronsteers Some recent conversations with community members have sparked the idea for a performance benchmarking/troubleshooting section. Possible rough content outline 1. Start by listing all the possible areas of slowdown: CLI startup, tap throughput, tap rate limits, target throughput, target batch frequency, network latency, network bandwidth, RAM limitations and buffer overruns. 2. Benchmark a data movement noop such as

meltano config PLUGIN_NAME list

. 3. Benchmark the EL processing times from a single tap to a single target. 4. Benchmark the tap alone. This would echo STDOUT to file(s) or to a null buffer. 5. Benchmark the target alone. This would take data from pre-generated file(s) to isolate target performance from the performance of the tap. Iterative approaches We could absolutely start this off with some very basic instructions on just a few of the above bullets. A good first iteration would at minimum give basic instructions to run the tap and target separately, since that is the first step in identifying where the EL pipeline is being throttled. Related discussions • meltano/sdk#906 from @BuzzCutNorman • #6597 from @dwallace0723 • meltano/sdk#9 which would benefit from before-and-after benchmark data. cc @edgarrmondragon, @visch, @pandemicsyn, @afolson, @tayloramurphy meltano/meltano