Parallelism Per Request

So I’m running Elasticsearch on a single 72 core server. I’ve created a very large 72 shard index, and would like to run searches that leverage all 72 cores. However, when I run a search request, according to the /_cat/thread_pool only 5 threads are being used per request. I’m figuring that I’m doing something dumb, but it’d be great if anyone has any insight on how I could simultaneously leverage more of the cores. Thanks!

Just to make sure, how did you figure out that only 5 threads are used?

The more important question to me is, do you really want to optimize for the one query use case, where only one user in parallel fires of a query? It's likely that you will have more than one user using your system in parallel, so this is what you should cater for from my outside perspective. This also means you are probably fine with just having one shard and many parallel queries by parallel users. Then you could see how this plays out from a search query time and increase the number of shards.

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.