Choosing tracks that contain long-running tasks

[Sunt-ing (Ting Sun)]

Hi all! I need to benchmark on a track that contains some long-running (such as larger than 10s) tasks. Benchmarking one track after another is definitely an approach and I'm working on that, but there are so many tracks that benchmarking all of them may take a very long time.

May you help figure out the appropriate tracks that I need to benchmark? Thanks!

P.S. I run ES and do the whole benchmarking on my laptop (512GB SSD, 32GB DRAM).

[json (Jason Bryan)]

Hi Ting Sun, welcome to the channel, and thank you for posting. What type of long-running tasks do you have in mind? The Elastic logging track may have just what you are looking for as it contains long-running tasks for indexing and querying.

Benchmarking on your laptop is great for testing, especially when you've got some available memory. Be sure to give Seven Tips for Better Elasticsearch Benchmarks a read for tips on system setup and benchmarking methodology.

Thank you,
Jason

[Sunt-ing (Ting Sun)]

Hi Jason, thanks for your timely reply and warm welcome!

What type of long-running tasks do you have in mind?

As described in the blog A Short Emergency Response Guide for Elasticsearch written by Ugo Sangiorgi, task cancellation is an approach to save production on fire.
I implemented a simple script tool to somewhat automate this process and am currently trying to test my tool. So I need to have a workload that contains some long-running tasks which may be canceled by my tool.

The Elastic logging track may have just what you are looking for as it contains long-running tasks for indexing and querying.

Be sure to give Seven Tips for Better Elasticsearch Benchmarks a read for tips on system setup and benchmarking methodology.

Thanks for your recommendation! I will have a try as well as read the blog.