Data, Records Size, and Usage Limits
Algolia limits the size of a record for performance reasons. The limits depend on your plan - see our pricing page. We recommend a few techniques that can help you reformat and break up your records into smaller ones.
Records can’t go beyond a certain size limit of minified JSON:
- 10 KB for Pro, Starter, or Free accounts
- 20 KB for legacy (Essential and Plus)
- 100 KB or more for Enterprise plans.
You’ll get the
Record is too big error if you try to index a record that exceeds the limit.
You only need to worry about index size if your application runs on dedicated hardware (you are on an Enterprise plan). Though there is no strict upper limit to an index’s size, we recommend you keep indices smaller than 102GB. This ensures good performance.
This is 80% of the RAM capacity (128 GB) of dedicated servers, which leaves 20% of the RAM capacity to handle all of your indexing tasks. If the index size exceeds the 128 GB capacity, performance degrades severely: data is swapped back and forth between temporary and permanent memory, which is a very costly operation.
Keep in mind there is no limit on the number of records an index can have, there’s only a limit on the memory capacity of the hardware.
Indexing Usage Limits
Max indexing operations
Regarding pricing, we count the number of operations performed every month. If you hit the limit of your plan, we’ll charge you a fee for the extra operations you’ve performed, based on the over-quota pricing of your current plan.
Indexing rate limit
Algolia delays or rejects indexing operations whenever a server is overloaded. If Algolia determines that indexing operations will negatively impact search requests, it takes action to favor search over indexing. We call this the rate limit, put in place to protect the server’s search capacity.
A search operation is counted whenever a search is performed. In autocomplete and search-as-you-type implementations, a new search is performed on every keystroke. In case your search engine is querying N indices at each keystroke (a product index, a brand index, a categories index, etc.) then one keystroke will correspond to N operations.