Slow Elasticsearch Ingestion

I am new to ELK stack and I am currently testing on the performance of the Elasticsearch with 3 different configurations as follows:

  1. One source file: 1GB
  2. One source file: 3GB
  3. 3 source files: 1GB+1GB+1GB

It took me 3 to 4 hours to have the 3GB file ingested into Elasticsearch. It took me more than a day to finish 3 different 1GB files to be ingested. The speed was getting slower as the size of the index increases.

I have tried to increase the JVM heap size and it's getting slightly better, shortening the time by about one hour. All other settings are configured in the default settings.

I would like to know the additional ways on how to improve the ingestion speed on this? Thanks.

What is the specification of your cluster in terms of nodes and hardware used How are you infesting the data?

This is my current configuration of Elasticsearch. All are in default settings.

What is the specification of the host? What type of disk/storage do you have?

I am using VM here in an ESXI server so the hardware specification is as follows:

As I need this ELK for only one application so I didn't allocate high hard spec for this VM.

Indexing is I/O intensive so having fast storage is important for optimal indexing speed. I would recommend running ‘iostat -x’ on the node while indexing.

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