ES stack trace does not reflect altered thread pool configuration

Elastic Stack Elasticsearch
1

Hi
We have recently updated to ES 6.5.3 and changed the default configuration accordingly:

thread_pool.search.queue_size: 8000
thread_pool.search.size: 35

When querying this at runtime we get:

   "thread_pool" : {
          "search" : {
            "queue_size" : "8000",
            "size" : "35"
          }

...so the configuration modifications take effect. The problem is that when we increase the rate of our querying we get this stack trace:

Caused by: org.elasticsearch.common.util.concurrent.EsRejectedExecutionException: rejected execution 
of org.elasticsearch.common.util.concurrent.TimedRunnable@17f348a3 on 
QueueResizingEsThreadPoolExecutor[name = thu-match-es4/search, queue capacity = 1000, min queue 
capacity = 1000, max queue capacity = 1000, frame size = 2000, targeted response rate = 1s, task 
execution EWMA = 514.2micros, adjustment amount = 50, 
org.elasticsearch.common.util.concurrent.QueueResizingEsThreadPoolExecutor@31dbc67a[Running, 
pool size = 25, active threads = 20, queued tasks = 999, completed tasks = 2878951156]]

My question is, why does the stack trace not reflect our modified configuration? Is there a problem in 6.5.3 or have we set the configuration incorrectly?

Regards
Johan

2

Hi,

how many nodes does your cluster have? Did you apply this setting to all nodes?

Daniel

3

We have 4 nodes and the setting was applied to all nodes