Elastic Cloud

Elastic Cloud provides managed Elasticsearch — the most widely deployed search and log analytics engine. Key capabilities: full-text search with BM25 relevance scoring, structured aggregations (metrics, bucket, pipeline), vector similarity search (kNN with dense_vector fields), and log analytics at massive scale. ESQL (Elastic Query Language) provides a SQL-like syntax for analytics queries. For agents: the Elasticsearch REST API enables document indexing, multi-field search queries, aggregation-based analytics, and kNN vector search for RAG pipelines. Kibana provides visualization; the REST API enables programmatic data access. Self-hostable (open-source) or Elastic Cloud for managed. Confidence is docs-derived.

Elasticsearch REST API at {cluster_url}. Resources: indices, documents, search. POST /{index}/_doc indexes a document ({field1: value, field2: value}). GET /{index}/_search executes a search query ({query: {match: {title: "elasticsearch"}}, aggs: {...}, knn: {field: "embedding", query_vector: [...], k: 10, num_candidates: 100}}). GET /{index}/_doc/{id} retrieves a document. DELETE /{index}/_doc/{id} deletes a document. POST /{index}/_bulk for batch operations. POST /_msearch for multi-index search. POST /{index}/_update/{id} for partial document updates. Query DSL: match, term, bool, range, wildcard, fuzzy, nested, kNN. ESQL: POST /_query {query: "FROM my-index | WHERE field > 100 | STATS avg(price)"}.

API key auth (recommended): Authorization: ApiKey base64({id}:{api_key}). Keys from Kibana → Stack Management → API Keys. Bearer token: Authorization: Bearer {token}. Basic auth: Authorization: Basic base64({user}:{password}. HTTPS enforced. Kibana SSO: SAML/OIDC for enterprise. Elasticsearch security: index-level and field-level security via roles. Document-level security: role query restricts which documents a role can see. Key rotation: create new API key, revoke old via ES API.

Standard HTTP status codes with JSON error bodies. 400 for invalid query DSL (detailed parse error). 404 for missing index or document. 409 for version conflict (optimistic locking). Bulk indexing: partial success reported per-action in response. kNN search: num_candidates tuning for accuracy vs. speed tradeoff. Sharding: cluster rebalancing during index growth may temporarily slow queries. Elastic Cloud SLA: 99.9%+ for production tiers. Index lifecycle management (ILM) for automatic hot/warm/cold/delete transitions. Shard count tuning is important for performance — Elastic docs provide guidelines.

elastic.co/guide is comprehensive — Elasticsearch REST API reference, Query DSL guide, aggregations reference, kNN vector search guide, ESQL reference, and observability guides. Getting started: Elastic Cloud free 14-day trial (no card), deploy cluster, first index + search in under 10 minutes. Official client libraries: Python (elasticsearch-py), JavaScript (@elastic/elasticsearch), Go, Ruby, Java, .NET, PHP. Kibana console for interactive Elasticsearch queries during development. Community via Elastic Discuss forum (discuss.elastic.co), GitHub, and Elastic Discord.