Advanced methods

Retry logic

To provide reliability, Algolia’s architecture is redundant.

Every application is hosted on at least three different servers (clusters).

As a developer, you don’t need to worry about these details, the API clients handle them for you:

• Use the dynamic Distributed Search Network to perform automatic load balancing between servers.
• The retry logic switches the targeted server whenever it detects that one of them is down or unreachable. A given request won’t fail unless all servers are down or unreachable at the same time.

Application-level errors such as invalid queries are still reported without retry.

Error handling

Requests can fail for two main reasons:

1. Network issues: the server couldn’t be reached, or didn’t answer within the timeout.
2. Application error: the server rejected the request.

For application errors, the API client reports:

• message: an error message indicating the cause of the error
• status: an HTTP status code indicating the type of error

For example:

1
2
3
4
{
  "message":"Invalid Application ID",
  "status":404
}