By enabling a default mode, the client adapts to pick an appropriate retry policy and network settings.
Without this setting, the SDK does not do any retries. In the past we have implemented our own retry logic. But I prefer this way because I trust the sdk to know which error categories should be retried. And it makes the aws clients consistent with our gcp/kafka clients, which also do retries under-the-hood.
The defaults mode is described in this AWS blog post.
By enabling a default mode, the client adapts to pick an appropriate retry policy and network settings.
Without this setting, the SDK does not do any retries. In the past we have implemented our own retry logic. But I prefer this way because I trust the sdk to know which error categories should be retried. And it makes the aws clients consistent with our gcp/kafka clients, which also do retries under-the-hood.