If you retrieve data by including variables in the query, the number will be limited to 2000.
If you download in Excel or CSV, you can download all the data of over 2000 items.
It is currently October 2023, but is it possible to change the limit on the number of displayed items in the free version of metabase?
Is it still only available for custom builds?
Trying to create time-based graphs is pointless with a 2000 line limit. It would seem more useful to limit the size of data not the amount of lines as that is what will affect the browser the most. If I have 1000 columns of maxchar data, 2000 lines is still gonna drag the browser to its knees. 10000 lines of 2 columns of numeric data won't.
It is kind a bit if a necro but this post is the latest related about this limit and I have found the fix.
If you are using the java jar version, simply change these two environment variables :
MB_AGGREGATED_QUERY_ROW_LIMIT=20000
MB_UNAGGREGATED_QUERY_ROW_LIMIT=10000
You can either set them with export or if you are using metabase as a service, place these exact two lines in /etc/default/metabase then do sudo systemctl daemon-reload then restart the metabase service.
Either using environment variables or -D options to the JVM works, it accomplishes the same thing.
-D may be more convenient if manipulating environment variables is difficult or impossible, or on Windows where environment vars work differently than their Unix counterparts.
Be aware that increasing the row limit will also increase Metabase’s and the browser’s memory use, increase database load and network traffic, and impact performance across the app. Set up Prometheus metric logging, watch for signs of memory stress in the JVM, and increase JVM memory if signs are spotted.
The Metabase browser chart code bogs down if there is a large number of points; it’s not as efficient as charting packages that draw on a canvas.