AWS OpenSearch MCP Server
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| AWS_SDK_LOAD_CONFIG | Yes | Set to '1' to enable AWS SDK to load config profiles from ~/.aws/config (required) | 1 |
| AWS_OPENSEARCH_MCP_CONFIG | Yes | Path to the configuration YAML file (required) | |
| AWS_OPENSEARCH_MCP_LOG_LEVEL | No | Logging level (optional, default INFO) | INFO |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": false
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
| experimental | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| list_aws_profilesA | List allowlisted AWS profiles, whether they exist locally, and allowed regions. |
| list_domainsB | List Amazon OpenSearch Service domains visible to an allowlisted AWS profile. |
| get_domain_configC | Read the managed-domain configuration, nodes, storage, networking and security posture. |
| get_cluster_healthC | Get OpenSearch cluster health at cluster, indices, or shards level. |
| get_cluster_statsC | Get high-level cluster, node, shard, document and storage statistics. |
| list_indicesB | List matching indices ordered by storage size with health, shards and document counts. |
| get_index_detailsB | Read settings, mappings, aliases and operational stats for an index or safe pattern. |
| search_indexC | Run a bounded read-only Query DSL search. Scripts and runtime mappings are blocked. |
| get_latest_documentsC | Return newest documents sorted by a date field, optionally restricted by lookback hours. |
| get_field_mappingB | Inspect one field across matching index mappings and report mapping type conflicts. |
| get_field_countC | Estimate mapped field counts and compare them with index.mapping.total_fields.limit. |
| get_shard_allocationB | List primary and replica shard states, nodes, sizes and unassigned reasons. |
| explain_unassigned_shardB | Explain the first unassigned shard, or a specific shard when all shard fields are given. |
| get_disk_allocationC | Get per-node disk allocation and shard counts ordered by disk percentage. |
| get_cluster_settingsC | Read transient, persistent, and optionally default cluster settings. |
| get_indexing_statsC | Get document, storage, indexing, search, merge, refresh and segment statistics. |
| get_pending_tasksC | List pending cluster-manager tasks and their queue wait time. |
| get_ingest_pipelinesC | Read all ingest pipeline definitions visible to the current AWS identity. |
| diagnose_clusterB | Generate an evidence-based read-only diagnosis for health, shards, disk and flood-stage blocks. |
| diagnose_timestampB | Diagnose stale data by checking date mappings, min/max values, and newest/oldest samples. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 20 tools
Each tool targets a distinct aspect of OpenSearch cluster management and diagnosis, such as health, shards, fields, ingest pipelines, and domain configuration. Even tools with overlapping themes (e.g., diagnose_cluster and explain_unassigned_shard) have clearly different scopes, minimizing confusion.
All tool names follow a consistent verb-noun pattern using snake_case (e.g., diagnose_cluster, get_cluster_health, list_indices). Verbs like 'get', 'list', 'diagnose', 'explain' are used predictably, making it easy for an agent to infer functionality.
With 20 tools, the set is well-scoped for a dedicated OpenSearch diagnostic server. It covers a wide range of inspection and diagnosis tasks without being overwhelming, and each tool serves a clear purpose.
The tool set comprehensively covers read-only diagnostics: cluster health, stats, settings, shard allocation, field mappings, ingest pipelines, and more. Minor gaps exist, such as lack of node-specific details or thread pool stats, but these are acceptable given the read-only and diagnostic focus.