Blank display when I add a filter on discover page since update opensearch from 3.1 to 3.3

This doesn't appear to be a documented issue specifically related to upgrading from OpenSearch 3.1 to 3.3. However, there are a few things you can check to troubleshoot this blank display issue:

1. Check your browser's developer console for any error messages that might indicate what's happening when you apply the filter.

2. Verify that your IAM permissions are correctly configured. The Discover page in OpenSearch Dashboards requires both ESHttpGet and ESHttpPost permissions to function properly. Even for read-only operations, POST requests are used because they allow for more complex query structures and larger payloads than GET requests. If your permissions are not properly configured, you might see issues like blank pages or loading indicators that keep spinning.

3. Check if any CloudWatch metrics or alarms were affected by the upgrade. When upgrading between OpenSearch versions, some metrics can change, which might break existing alarms or monitoring. This could potentially impact the functionality of your OpenSearch domain.

4. Consider checking if there were any incompatible index settings that weren't addressed before the upgrade. While this is more commonly an issue when upgrading from 2.x to 3.1, it's possible that similar compatibility issues exist between 3.1 and 3.3.

If the issue persists, you might want to contact AWS Support for more specific assistance with your OpenSearch Service domain configuration.
Sources
OpenSearch UI dashboard's Discover page requires ESHttpPost permission even when you only want to grant read access | AWS re:Post
Amazon OpenSearch Service rename - Summary of changes - Amazon OpenSearch Service
Upgrading Amazon OpenSearch Service domains - Amazon OpenSearch Service

I had no problem in 3.1 version. I'm admin in the cluster so I have full access.

In developer console, I have the following error when I set a filter.

I don't have this error when I add filter on our production Opensearch Cluster.

Thanks in advance

I clean up my browser caches and every thing works fine now. We can close this post.

Hey,

Hope you're keeping well.

This behavior after upgrading to OpenSearch Dashboards 3.3 can sometimes be caused by index mapping changes or incompatible saved objects from the earlier version. I’d suggest checking the browser developer console for the exact API call failing when the filter is applied, then compare that request against the same action in your production cluster. Also verify that the index pattern in Stack Management > Index Patterns is still valid and that the field you’re filtering on exists with the correct type in the mapping. If the issue persists, try exporting your saved objects, deleting and recreating the index pattern, and re-importing, as this can clear stale metadata from the upgrade.

Thanks and regards,
Taz

Hey,

Hope you're keeping well.

This looks like a Dashboards UI or saved object compatibility issue introduced in 3.3. Check if the index pattern in Discover is still valid and that the field you’re filtering on exists in the mapping without type changes. You can try re-saving the index pattern in OpenSearch Dashboards > Stack Management > Index Patterns, or export/reimport saved objects to refresh them. Also clear your browser cache and test in an incognito window to rule out stale assets. If the error persists, capture the failing API call from the browser console and open an AWS Support ticket, as there have been some UI regressions reported in recent 3.x updates.

Thanks and regards,
Taz