[BUG] incredibly slow dashboards since > v2.5.0

[mshwery]

What is the bug?

At some version after 2.5.0 of dashboards, the dashboard became painfully slow. I cannot pinpoint the exact version because of how AWS OpenSearch upgrades get applied. We are currently on the latest patch of 2.11

How can one reproduce the bug?

Install opensearch dashboards via docker-compose, navigate to the index management page, click around, look at mappings... OR open the console in the dashboard and try typing anything or hitting enter on a query.

What is the expected behavior?

Not for keyboard input or clicks to hang. They should be responsive to input throughout the dashboard.

What is your host/environment?

AWS OpenSearch

Do you have any screenshots?

I took a random sampling of the javascript performance itself, and that may be part of the bottleneck. Check out these long tasks:

Do you have any additional context?

We have ~1200 indices. Some have a large number of mappings (~5k). I suspect there is something really slowing down in JavaScript in the console. Performance is worst there. But it's also slower than before everywhere else too, including the index management pages, the mappings page, etc.

[ananzh]

@mshwery from the screenshot, this is for Discover only right?

[mshwery]

@mshwery from the screenshot, this is for Discover only right?

It seems more global, it also affects other pages like index management pages, aliases. Those are the ones I frequently check so I'm not sure if it's more widespread.

[ananzh]

@mshwery we will reproduce this and will reach out to you.

[rayangou]

@ananzh Any updates on this?. We are also experiencing this issue