Aggregation queries compute summary values (averages, totals, extremes) across time-series data points instead of returning individual rows. Use these when you need calculated metrics rather than raw data — for example, the average voltage over the last hour or the total energy produced per machine.
A downsampling query reduces the density of time-series data by dividing a time range into equal-sized buckets and applying an aggregation function to each bucket. This is essential for visualizing large datasets on dashboards and charts — instead of transferring thousands of raw data points, you get a manageable number of summarized values.
Persisted queries are saved query definitions stored as runtime entities. Instead of specifying all query parameters every time, you create a query definition once and execute it by its rtId. This is useful for dashboard widgets, recurring reports, and any scenario where the same query needs to be executed repeatedly.
A simple query retrieves raw time-series rows from CrateDB. You choose which columns to return and can apply filters, sorting, and time range restrictions. This is the most fundamental stream data query type — use it when you need individual data points rather than aggregated summaries.