Setting the file. One moment.
Chapter 45 · Amazon Opensearch Service
Subchapter 45.47
assets/elasticsearch-index-template-skeleton.mdMarkdown4 KBView on GitHub
Use this when the source is Elasticsearch or OpenSearch. Most ES/OS mappings carry over 1:1; this skeleton is the audit target for the handful of constructs that need action (see the table in ). For Solr sources use instead.
Migration Assistant for Amazon OpenSearch Service does this for you. Historical Data Migration’s metadata-migration phase translates the source mappings + index templates into OpenSearch-compatible form (stripping
_type, convertingdense_vector→knn_vector,flattened→flat_object) and reindexes documents. This skeleton is for auditing Migration Assistant for Amazon OpenSearch Service’s output and for the rare override — NOT a manual step in the migration plan.
{
"index_patterns": ["<index-name>-*"],
"template": {
"settings": {
"number_of_shards": "<from Step 5 sizing>",
"number_of_replicas": 1,
"refresh_interval": "30s",
"analysis": {
"analyzer": {
"<custom_analyzer>": {
"type": "custom",
"tokenizer": "<tokenizer>",
"filter": ["lowercase", "<filter>"]
}
}
}
},
"mappings": {
"properties": {
"<keyword_field>": { "type": "keyword" },
"<text_field>": {
"type": "text",
"fields": { "keyword": { "type": "keyword", "ignore_above": 256 } }
},
"<int_field>": { "type": "integer" },
"<date_field>": { "type": "date", "format": "strict_date_optional_time||epoch_millis" },
"<geo_field>": { "type": "geo_point" },
"<vector_field>": {
"type": "knn_vector",
"dimension": "<dim>",
"method": { "name": "hnsw", "engine": "faiss", "space_type": "l2" }
}
}
}
}
}index_patterns matches the target index / alias name.number_of_shards / number_of_replicas come from Step 5 (Estimate Sizing); refresh_interval defaults to 30s for prod per sizing.md, not 1s._type removed. Multi-type (ES 6.x) or _doc-placeholder (ES 7.x) mappings are flattened — types do not exist in OpenSearch (nugget #9).fielddata: true stripped from text fields and replaced with a .keyword subfield + doc_values (nugget #8) or the node OOMs on first aggregation.dense_vector → knn_vector with an explicit method/engine chosen per the k-NN engine table in vector-knn.md; recall validated against source. [verify] the current default engine for the target version.flattened → flat_object.runtime mapping equivalent); reindex required._source: {enabled: false} indexes are migrated via Migration Assistant for Amazon OpenSearch Service Historical Data Migration only (nugget #22), and _source is re-enabled on the target.alias type) carry over unchanged.analysis; filter order preserved (lowercase before synonym_graph/stop).EncryptionAtRestOptions); node-to-node encryption (NodeToNodeEncryptionOptions); HTTPS enforced (EnforceHTTPS: true, TLSSecurityPolicy: Policy-Min-TLS-1-2-2019-07); fine-grained access control with IAM/SAML/OIDC; access policy scoped by principal and source ARN/account. You MUST NOT use 0.0.0.0/0.CustomEndpoint.CertificateArn). You MUST NOT use a self-managed certificate.You MUST embed the filled-in template in section 2. Schema / Mapping of the report. You MUST cite the source _mapping field name for each non-trivial mapping decision in the table.