Constant timeouts querying Packetbeat indices

[Wanderer]

We're running one ES/Kibana server and pulling in Filebeat, Metricbeat and Packetbeat data. The Filebeat and Metricbeat indices are a few hundred MB and run fine but Packetbeat indicies are several GB. When we include Packetbeat indices in Discover queries we get timeouts and the CPU spikes for long periods. I can increase the Kibana timeout to mask the problem. I've removed flow data to decrease the size of the indices but even searching for a 30 minute window causes timeouts in Kibana.

Error: Gateway Timeout
at respond (https://es-data.veerotech.net/bundles/vendors.bundle.js?v=16588:111:161556)
at checkRespForFailure (https://es-data.veerotech.net/bundles/vendors.bundle.js?v=16588:111:160796)
at https://servername/bundles/vendors.bundle.js?v=16588:105:285566
at processQueue (https://servername/bundles/vendors.bundle.js?v=16588:58:132456)
at https://servername/bundles/vendors.bundle.js?v=16588:58:133349
at Scope.$digest (https://servername/bundles/vendors.bundle.js?v=16588:58:144239)
at Scope.$apply (https://servername/bundles/vendors.bundle.js?v=16588:58:147018)
at done (https://servername/bundles/vendors.bundle.js?v=16588:58:100026)
at completeRequest (https://servername/bundles/vendors.bundle.js?v=16588:58:104697)
at XMLHttpRequest.xhr.onload (https://servername/bundles/vendors.bundle.js?v=16588:58:105435)

[ruflin]

What happens if you run the queries directly against Elasticsearch? What is your Elasticsearch cluster size? I'm wondering if you hit an issue on the Elasticsearch or Kibana side.

[Wanderer]

We are rookies at this . Can you point me to where I can find how to do this in the docs please?

[ruflin]

So basically the question is how many Elasticsearch nodes have you deployed with how much memory? And what is the size of your indices? https://www.elastic.co/guide/en/elasticsearch/reference/current/indices-stats.html