Capacity alerts warn you when something is about to go wrong on your servers. This is useful in helping you anticipate and avoid serious troubles. Still, it can also reveal a need for optimization, such as in how you define your settings and search parameters, or how you structure and push your data.
There are many considerations to take into account, and we can help you find solutions.
Only Enterprise customers with dedicated servers can receive capacity alerts.
What do we monitor?
We monitor several kinds of server-side metrics, so we can alert you when one reaches full capacity.
This is for the combined total size of all indices on the clusters. If the size of your indices reach the server’s limits, there’s a risk of disrupting the service.
This is the available CPU for search queries. Some queries are delayed—or eventually dropped—if you consume 100% of the search capacity.
Average response time
This is the average time for the engine to process a query. You can fix slow queries by optimizing query parameters or the index’s configuration.
Oldest queued job
This is the delayed indexing queue on your servers. It usually happens when the engine receives a large volume of complex indexing operations.
How to configure and subscribe to capacity alerts
Every collaborator can subscribe to capacity alerts for each application. You can set this up on your Algolia dashboard, from the Alerts tab.
You can configure a threshold for each metric. However, the default values should be sufficient for most applications, so we recommend not changing them.
If you need help with capacity alerts, feel free to reach out to your Solutions Engineer, CSM, or send us an email at email@example.com.