With automatic indexing turned on, the indexing happens asynchronously. It may take some time for your data to be updated.
It is possible for the indexing queue to crash under certain circumstances. Below is a list of the most common problems you can encounter.
Network error
Network errors are typically resolved automatically thanks to the queueing process. If the network error persists over multiple queue runs, there may be a problem with the underlying infrastructure of the Magento application. When in doubt: if the problem lies with the Algolia infrastructure, please check the status page for any issues.
Memory error
Magento has some known memory leaks which cause problems when memory usage increases. When sending records to Algolia that exceed 10K in size, memory usage increases. Algolia doesn’t only suggest, but in fact requires record sizes to not exceed 10K for performance reasons.
Record size
If the total size of the index exceeds this limit, it’s recommended to turn on automatic indexing through the queue. The cron job that runs the indexing will break up large indices into chunks within the size limit. When indexing without the cron job, or with the EMPTY_QUEUE=1 argument in the job, the index size isn’t checked.
Failed (Re)indexing jobs
To see if any of the jobs that were created have failed, please have a look in the algoliasearch_queue table. If a job fails, it needs to be restarted. This will continue the indexing job from where it stopped, so it won’t run it in its entirety again. If the queue is enabled, the restart will happen automatically. The indexing job can be restarted manually by pressing the restart link. This can be found by clicking the index on the Magento Dashboard.