# How ECE Kibana balances requests

**URL:** <https://discuss.elastic.co/t/how-ece-kibana-balances-requests/167885>\
**Category:** Elastic Cloud Enterprise (ECE)\
**Created:** [February 11, 2019, 3:34pm UTC](https://discuss.elastic.co/t/how-ece-kibana-balances-requests/167885 "2019-02-11T15:34:19Z")\
**Posts on this page:** 4\
**Page:** 1

<div class="post-metadata">

**Author:** ![osykora](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/osykora/32/41464_2.png) [@osykora](https://discuss.elastic.co/u/osykora)\
**Post date:** [February 11, 2019, 3:34pm UTC](https://discuss.elastic.co/t/how-ece-kibana-balances-requests/167885/1 "2019-02-11T15:34:19Z")

</div>

Hi,

how is Kibana that is part of an ece cluster balancing requests inside the cluster? I think showing an example will be best:

A cluster with three zones A, B, C. Each zone having 2 ES nodes + 1 Kibana.

1. When a request reaches Kibana in zone A, is it then calling one of two ES nodes in the same zone or can it call any ES node in any zone?
2. This relates to the question above, but if one or both ES nodes fail in zone A, will Kibana in the zone A still be able to handle requests correctly?

Appreciate your help

---

<div class="post-metadata">

**Author:** ![Alex\_Piggott](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/alex_piggott/32/11053_2.png) [@Alex\_Piggott](https://discuss.elastic.co/u/Alex_Piggott)\
**Post date:** [February 11, 2019, 6:46pm UTC](https://discuss.elastic.co/t/how-ece-kibana-balances-requests/167885/2 "2019-02-11T18:46:31Z")

</div>

> 1. When a request reaches Kibana in zone A, is it then calling one of two ES nodes in the same zone or can it call any ES node in any zone?

It will randomly `(*)` choose between nodes in the same zone, and _can_ go x-zone if there are no available nodes in the same zone (choosing randomly between other zones, and other nodes in the same zone `(*)`)

`(*)` Actually this is not entirely random, eg it try to route to nodes of the most sensible type, eg client (read) or ingest (write) nodes first, then data nodes, then master nodes etc

As a result of the above, it will continue to function even if all nodes in one zone are unavailable

Alex

---

<div class="post-metadata">

**Author:** ![osykora](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/osykora/32/41464_2.png) [@osykora](https://discuss.elastic.co/u/osykora)\
**Post date:** [February 12, 2019, 8:59am UTC](https://discuss.elastic.co/t/how-ece-kibana-balances-requests/167885/3 "2019-02-12T08:59:02Z")

</div>

Thanks Alex

---

<div class="post-metadata">

**Author:** ![system](https://us1.discourse-cdn.com/elastic/original/3X/1/a/1ac57faf039f6b580b3f104ef42a2a89e41014de.png) [@system](https://discuss.elastic.co/u/system)\
**Post date:** [February 26, 2019, 9:09am UTC](https://discuss.elastic.co/t/how-ece-kibana-balances-requests/167885/4 "2019-02-26T09:09:13Z")

</div>

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.