Elastisearch error queued tasks

I have some problems regarding to display graphs using Elastisearch. The error is that my queued tasks are above 1000. Is there some way to get around this?
I can get passed this by going in on each graph and press on Query Inspector, but this is quite time consuming. Any other way around this?

That is likely an Elasticsearch issue. When you are viewing the Query Inspector, only the graph that you are making changes to is querying your ES cluster, so there are fewer inflight queries. When viewing the dashboard, all the visible graphs are querying the ES cluster at the same time.

You can increase the search queue to work around that issue, but you will need resources in your cluster to handle it as well. Our cluster is running with the following:

  • thread_pool.index.size: “32”
  • thread_pool.index.queue_size: “5000”
  • thread_pool.bulk.size: “32”
  • thread_pool.bulk.queue_size: “5000”
1 Like