Details
-
Improvement
-
Status: Open
-
Minor
-
Resolution: Unresolved
-
1.40.0
-
None
-
None
Description
Currently we have the following retry configuration for elastichsearch:
public static final String BULK_RETRIES = "bulkRetries"; public static final int BULK_RETRIES_DEFAULT = 3; public static final String BULK_RETRIES_BACKOFF = "bulkRetriesBackoff"; public static final long BULK_RETRIES_BACKOFF_DEFAULT = 200;
This does not really protect against network hiccups which can take longer. I would suggest either 5 retries with 500 backoff or 3 retries with 1s backoff.