© 2026 Hashnode
If you are building a SaaS product on top of Kusto, you likely have one massive cluster where all your customers' data lives together in the same tables (e.g., Logs, Events). You want to give your customers the power to run KQL queries, but you can't...

Calculating rolling statistics (like a moving median or the 95th percentile over the last 10 days) used to be a pain in KQL. You either had to perform expensive self-joins or chain multiple prev() functions together, which is ugly and hard to maintai...
