v.25.7Improvement
Cases, we need to have multiple dimensions
In some cases, we need to have multiple dimensions to our metrics. For example, counting failed merges or mutations by error codes rather than having a single counter. Introducesystem.dimensional_metrics, which does precisely that and adds the first dimensional metric calledfailed_merges. #83030 (Miсhael Stetsyuk).
Why it matters
This feature adds the ability to track metrics segmented by multiple dimensions, such as counting failed merges or mutations by their error codes, rather than aggregating everything into a single counter. This enhances observability and troubleshooting by providing more granular metric details.How to use it
Query the newsystem.dimensional_metrics table to access dimensional metrics like failed_merges. Use standard SQL SELECT statements to filter and analyze metrics by dimensions such as error codes.