this might not be related. but I had this unusal slow opration of kibana for everything and anything on it.
Turn out that they have turn on fleet by default. ( I believe that was from 7.12+)
and that means your kibana is trying to go out and update fleet and if you don't have internet connection it is just spinning for long time.
disable fleet on kibana
xpack.fleet.enabled: false
Again this might not be the case in your situation.
Hi Sachin,
I'm afraid adding fleet setting didn't help.
It is clear to me the reason Discover view in 7.13.2 is slower than in 7.4.0 is because the size of the response is much bigger, since it doesn't use 'gzip' for content encoding.
Is this the expected behavior now? Or is there a way to improve response times in Discover view?
To improve performance of Discover View I have updated setting 'Number of rows' (discover:sampleSize) in Advance Settings. Default value is 500 - new value 50.
This reduce response time from 31 seconds to only 7 seconds
I suppose time reduced will depend on the size of the records - in our case documents are quite large.
Do you have many multi fields in your mappings? could you record and share a Chrome performance profile when loading Discover?
Many thx!
Best,
Matthias
Hi Matthias,
Yes, we do have many multi fields and we are working to reduce these.
Here you have details from Chrome when Number of Rows is set to 500:
Hi Matthias,
There is a small improvement after setting discover:searchFieldsFromSource to on, as response size is smaller. However, discover view is still much slower than version 7.4.0.
For the moment we'll keep discover:sampleSize set to 50 till 7.14 is released. This give us decent performance.
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.