How to Optimize Database Queries and Performance in Bubble

Open your Bubble editor and click the Logs tab in the left sidebar. Look at the Server logs section — each entry shows the workflow or page load action, the time it took, and the workload units consumed. Sort by duration to find the slowest operations. Next, preview your app and activate the Debugger by clicking the bug icon in the bottom-right. Step through page loads and workflow actions to see which searches take the longest. Also check Settings → App Metrics to see which pages and actions consume the most WUs over time. Write down the top 3-5 slowest searches to prioritize your optimization work.

Find any Do a Search For that uses the :filtered operator after the search. This operator downloads all matching records to the browser first, then filters them client-side — extremely slow for large datasets. Move those filter conditions into the search constraints instead. Click on your Do a Search For, and in the constraints area, add conditions that match what your :filtered was checking. For example, change 'Do a Search for Products:filtered (Category = Shoes)' to 'Do a Search for Products' with a constraint 'Category = Shoes'. Database constraints run on the server and use indexes, making them dramatically faster.

Open your Design tab and check your Repeating Groups. Click on elements inside each cell and look for any that have their own Do a Search For as a data source. A search inside a repeating group cell runs once per visible row — if you show 20 items and each cell has a search, that is 20 separate database queries on every page load. Fix this by pre-computing the data. For example, if each Product cell searches for its review count, add a 'review_count' number field to the Product Data Type. Use a backend workflow to update this count whenever a review is created or deleted. Then display the field directly instead of running a search.

Select your Repeating Group element and check its layout. If it shows all records at once with no fixed number of rows, it forces Bubble to load every matching record. Change the Repeating Group to a fixed number of rows — 10 to 20 is optimal. Add pagination controls below the group: a Next button and a Previous button. Create a custom state on the page called 'current_page' of type number with a default of 1. Set the Repeating Group's data source to Do a Search for your type with items from calculated offset. The Next button increments current_page and Previous decrements it.

If the same search is used by multiple elements on a page, running it multiple times wastes WUs. Instead, store the result once. Create a custom state on your page element: click the page in the Element Tree, then click the 'i' icon to add a custom state. Name it 'cached_products' with type Product and check 'is a list'. In a Page is loaded workflow, add a Set State action: set page's cached_products to Do a Search for Products with your constraints. Then point your Repeating Group's data source to page's cached_products and your count text to page's cached_products:count. Now one search serves multiple elements.

Identify any searches that retrieve data that rarely changes — categories, status labels, countries, or configuration values. These are better stored as Option Sets because they are cached as part of your app's code and require zero database queries and zero WUs. Go to the Data tab, click Option sets, and create a new Option Set. Add options and attributes for any extra data like icon or display_order. Then replace your search-based dropdowns and filters with Option Set references.