Search-queued-tasks more than 1000


(minhua) #1

I found that the search-queued-tasks was usually more than 1000.

There are several visualizes in my dashboard. Each of them set an aggregation at least. I have no idea what caused this issue.

Is this situation normal?


(Christian Dahlqvist) #2

How many shards are you typically searching with these dashboards?


(minhua) #3

There are six visualizes in my dashboard. Each of them set an aggregation at least.

The dashboard monitors a special index by daily and refreshes the dashboard page automatically by each 5 seconds .

The data of this index is split by daily. So the index pattern is like 'indexname.yyyy.mm.dd'.
Each index in elasticsearch has 5 shards.

The environment of my elasticsearch is non-clustered. Because it has only one node.


(Christian Dahlqvist) #4

How long does the dashboard take to reload? What is the specification of the node? How many days is the dashboard querying?


(minhua) #5

The dashboard take less than 1 second to reload.

As usual, the dashboard load data by current date.

Please check the specification of the node below:

4 core cpu
16G ram
window 2008(a virtual machine hosted on aliyun.com)

I set 8G ram by config the panel of elasticsearch-service to start up the elasticsearch service.

I have set the 'thread_pool.search.queue_size' up to 3000 this morning.

It seems that the kibana works well.

But I still have a little concern about the reason why the number of 'search-queued-tasks' is so high.


(system) #6

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.