There are several reasons why you might see a records usage spike in the API Monitoring section of your dashboard:
1. You are updating your records using replaceAllObjects.
This method creates a temporary index then swaps it in
During the build, you can effectively have two copies of your dataset (old + temp index), which can significantly inflate your record count.
When calculating monthly usage, we ignore the 3 days with the highest number of records meaning you can typically perform a full reindex using
replaceAllObjectson 3 days per month, without incurring any additional usage costs. More frequent reindexes can start counting toward usage.Updating individual objects as and when required, and only performing full reindexes a few times a month is usually the best approach to keep your data up to date.
2. You are using Standard Replicas
Standard Replicas copy the primary index’s records, so they increase total record usage.
3. You add records to a Primary Index that has Standard Replicas
If you “replicate” one of your indices and then add new records to the primary index, the record count of any associated Standard Replica indices will also increase
4. You are using the Magento extension + indexing queue
Using Magento's indexing queue to send updates to products, categories, pages, and other data to Algolia’s servers asynchronously creates temporary indices.
These temporary indices can produce temporary record fluctuations.
5. You have subscribed an Algolia Committed plan
On some committed plans, the number of records is being calculated for the entire year instead of just one month, which can make the count look unexpectedly high.