EU alternative · Amazon OpenSearch

European alternative to Amazon OpenSearch

Full-text search is a search index under resources.search_indexes. Sources point at tables or buckets in the same project. An OpenSearch domain, its cluster, and its index mappings are not imported.

What you declare for search

The index has a name and a list of sources. A table source names the table and the fields to index. A bucket source can limit itself to a prefix. index_type text is full-text search. index_type vector is a separate index for embeddings, not a second cluster.

Index storage is 0.04 credits per GB-hour. Query compute is 0.01 credits per GiB-second. The index is reviewed in the same plan as the table it reads.

What stays behind in OpenSearch

An OpenSearch or Elasticsearch domain is not parsed. Instance types, dedicated masters, shard counts, analyzers, and access policies do not become fields on the search index.

You list the records and the fields. You do not export the domain and expect that JSON to apply. Retrieval stays next to the tables and buckets already in the project.

Amazon OpenSearch vs KERNSTACK on European Bare Metal

Dimension Amazon OpenSearch KERNSTACK Search
Operator AWS account KERNSTACK project
Manifest Amazon OpenSearch console or API resources.search_indexes
Region AWS region list European bare metal
Bill Amazon OpenSearch line item One project credit balance

Frequently asked questions

No. You declare resources.search_indexes with sources and fields. Cluster size, shards, and access policies are not copied.

Tables and buckets in the same project. A table source lists the fields to index.

On a search index with index_type vector. It is not an OpenSearch k-NN plugin configuration brought across.

Ship your stack on the Software Lifecycle Cloud.

Closed preview on dedicated European infrastructure. Bring a repository or start from a blueprint.