Describe the feature:

It was observed in the field that Kibana Upgrade Assistant does not take into account the max headroom settings (c.f Max headroom settings).

If a reindexing is required, Kibana performs some checks to verify whether there any nodes above the low watermark (c.f code) and displays a warning accordingly (e.g screenshot below):

The code only checks the disk usage against the cluster.routing.allocation.disk.watermark.low cluster setting.

This warning could be confusing for users who have configured max headroom settings. Even if some nodes are above the low watermark, there could be still plenty of space for reindexing if the nodes have not yet reached the max headroom values.

Since Elasticsearch 8.15, the nodes stats API return the effective watermarks in the cluster (c.f Support effective watermark thresholds in node stats API and Nodes stats API) with the following data in the fs object:

• low_watermark_free_space / low_watermark_free_space_in_bytes
• high_watermark_free_space / high_watermark_free_space_in_bytes
• flood_stage_free_space / flood_stage_free_space_in_bytes

The Kibana Upgrade Assistant could therefore use the effective watermarks reported by Elasticsearch to provide a more accurate disk usage assessment and whether reindexing could potentially fail.

Describe a specific use case for the feature:

• Users that have configured max headroom settings and preparing to upgrade to 9.x