There is no mechanism to failover to search from DynamoDB or any other resource about fhir-works-on-aws-search-es HOT 3 CLOSED

By spreading your ES master nodes and data nodes into three availability zones it should minimize the chance of ES being down and therefore unreachable by your ddbToEs lambda. Link here

However, it is important that we implement mechanisms to ensure that all records from DDB are written to ES in case of a network error or other issues in which ddbToEs fail to write to ES in the first try.
Possible solutions include writing failed requests to a dead letter queue to be processed at a later time. Or to implement exponential back off time for writing the failed requests to ES.

I'll add a ticket to our back log for this story (FHIR-385).

from fhir-works-on-aws-search-es.

This issue is the same as the persistence layer issue: awslabs/fhir-works-on-aws-persistence-ddb#18

Leaving both open as it affects persistence and search components

from fhir-works-on-aws-search-es.

We are now logging failures into a DLQ and improved logging to identify the problematic resources

from fhir-works-on-aws-search-es.