Batching Control
Backoff/Retry Settings
In situations where the destination tool is overloaded and/or returns HTTP 5xx or 4xx error codes, exponential backoff kicks in to prevent overloading the destination tool's API.
- Max Batch Size: Maximum number of records to send in a single POST to the destination API
- Typically: 1000, but this depends on the destination’s ability to buffer the incoming data
- Max Batch Delay: the maximum time to wait before building a batch and processing.
- Typically: 5 seconds
- Base Retry Delay: The Initial delay time. i.e for 2 seconds, the retries would proceed 2, 4, 8, 16, ... seconds later.
- Max Intra-Retry Delay: Hard limit for the above retry delay.
- For example: If set to 60 seconds, retries would be performed again 2, 4, 8, 16, 32, 32, 32... seconds later.
- Max Retry Duration: Total time for retrying delivery. For REST destinations, a delivery is a batch of events in one REST POST request.
- For example: If set to 8 hours, then requests would be made based on the above settings for 2+4+8+16+32+32+32... until 8 hours in total duration is reached, then no subsequent attempts are made to retry sending the data.
- The assumption here is that if there is an outage or other issue with the destination tool, then recovery will take less than "Max Retry Duration" to complete.
Tip: For faster searching, add an asterisk to the end of your partial query. Example: cert*