- 최신
- 최다 투표
- 가장 많은 댓글
In AWS Managed OpenSearch Service, you can’t directly allocate separate resources for these tasks.
Instead, you can optimize performance by choosing the right instance types and sizes, optimizing shard allocation, defining index mappings and settings, using parallel execution and scaling, and enabling cross-cluster search and replication.
- Understand that we can always provide more resources to the cluster overall.
- How can shard allocation, index mapping help with resource contention between indexing and searching?
- Re. cross-cluster search replication, are you suggesting a follower cluster as a read replica used only for searching?
Hello,
As you rightly pointed out OpenSearch serverless allows you to segregate the Search and Indexing workloads. But in OpenSearch service you can't allocate the resources for Search and Indexing requests.
Please follow below guide to improve indexing and search requests:
- https://repost.aws/knowledge-center/opensearch-indexing-performance
- https://repost.aws/knowledge-center/opensearch-latency-spikes
If you're running indexing heavy operational analytics workloads such as log analytics, observability, or security analytics, you can benefit from the improved performance and compute efficiency of OR1 instances. In addition, the automatic data recovery offered by OR1 instances improves the overall reliability of your domain. For more info, please refer the blog
