-
Book Overview & Buying
-
Table Of Contents
The Definitive Guide to OpenSearch
By :
When you send a query to OpenSearch, the query passes through a number of processing steps before OpenSearch sends back the result.
Figure 5.1: Query processing chain
As shown in Figure 5.1, first, your search request lands on a coordinating node. Depending on how you set up your cluster and how you send your requests to the cluster, the coordinating node might be a dedicated coordinator node, a data node, an ML node, or a cluster manager node. Dedicated coordinator nodes are a good idea for high-scale workloads, but usually, you’ll use data nodes as request coordinators (see Chapters 13 and 14 for more on scaling, including dedicated coordinators).
Request queues and thread pools
In order to manage the work that each node performs, OpenSearch creates thread pools with associated queues to hold tasks for the threads. You can use the _cat/thread_pool API call to see all of the queues. (Note: the response shows a snapshot...