Skip to content

Enterprise Quickstart

Get up and running with MCP for Redis Enterprise in 5 minutes.

Prerequisites

  • Redis Enterprise cluster with REST API access (port 9443)
  • Admin credentials for your cluster(s)
  • redisctl installed (Installation Guide)

Step 1: Create a Profile

redisctl profile set my-cluster \
  --type enterprise \
  --url https://your-cluster:9443 \
  --username admin@redis.local \
  --password YOUR_PASSWORD \
  --insecure  # if using self-signed certificates

Verify it works:

redisctl -p my-cluster enterprise cluster get

Multiple Clusters

A key advantage of redisctl is multi-cluster management. Add profiles for each cluster:

# West region cluster
redisctl profile set cluster-west \
  --type enterprise \
  --url https://west.example.com:9443 \
  --username admin@redis.local \
  --password YOUR_PASSWORD \
  --insecure

# East region cluster
redisctl profile set cluster-east \
  --type enterprise \
  --url https://east.example.com:9443 \
  --username admin@redis.local \
  --password YOUR_PASSWORD \
  --insecure

# Central region cluster
redisctl profile set cluster-central \
  --type enterprise \
  --url https://central.example.com:9443 \
  --username admin@redis.local \
  --password YOUR_PASSWORD \
  --insecure

Step 2: Configure Your AI Assistant

Add to your MCP configuration (read-only by default - explore safely without modifying anything):

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "redisctl": {
      "command": "redisctl-mcp",
      "args": ["--profile", "my-cluster"]
    }
  }
}

Add to .mcp.json in your project:

{
  "mcpServers": {
    "redisctl": {
      "command": "redisctl-mcp",
      "args": ["--profile", "my-cluster"]
    }
  }
}

macOS: ~/.cursor/mcp.json

{
  "mcpServers": {
    "redisctl": {
      "command": "redisctl-mcp",
      "args": ["--profile", "my-cluster"]
    }
  }
}

Multi-Profile Configuration

For multi-cluster management, specify multiple profiles when starting the MCP server:

{
  "mcpServers": {
    "redisctl": {
      "command": "redisctl-mcp",
      "args": [
        "--profile", "cluster-west",
        "--profile", "cluster-east",
        "--profile", "cluster-central"
      ]
    }
  }
}

With multiple profiles configured, the AI can target specific clusters directly using the profile parameter on any tool call - no need to switch defaults.

Restart your AI assistant to load the new configuration.

Step 3: Try These Prompts

Start a conversation with your AI assistant and try these examples:

Cluster Overview

Prompt What it does
"Show me the cluster info" Cluster name, version, node count
"List all nodes in the cluster" Node IDs, addresses, status
"What's the cluster health?" Overall status and any issues

Database Management

Prompt What it does
"List all databases" All databases with memory and status
"Show database 1 details" Full configuration for a database
"Which databases have persistence enabled?" Filter by configuration
"What modules are loaded?" List available modules

License Management

Prompt What it does
"Check the license status" Expiration, type, features
"Show license usage" Shards and RAM used vs limits
"Is the license expiring soon?" Quick expiration check

Monitoring

Prompt What it does
"Are there any alerts?" Active cluster alerts
"Show cluster statistics" CPU, memory, throughput
"Get stats for database 1" Per-database metrics

Multi-Cluster Operations

Prompt What it does
"List my profiles" Shows all configured clusters
"Show cluster info for cluster-central" Query a specific cluster
"Check the license on cluster-west" Query with profile targeting
"List databases across all my clusters" Aggregate cross-cluster query

Optional: Enable Write Operations

By default, the MCP server runs in read-only mode - you can explore your clusters freely without risk of modifying anything. When you're ready for write operations, update your config:

{
  "mcpServers": {
    "redisctl": {
      "command": "redisctl-mcp",
      "args": ["--profile", "my-cluster", "--read-only=false"]
    }
  }
}

Write prompts to try:

Prompt What it does
"Create a 2GB database called cache-db" Creates a new database
"Enable maintenance mode" Prepares cluster for upgrades
"Update the cluster license with: LICENSE_STRING" Applies a new license

Example Conversation

You: What profiles do I have configured?

AI: You have 3 Enterprise profiles:

Name URL Default
cluster-west https://west:9443 Yes
cluster-east https://east:9443 No
cluster-central https://central:9443 No

You: Check the license on all my clusters

AI: I'll check each cluster:

cluster-west: Trial license, expires March 15, 2026 (38 days remaining) - Shards: 2/4 used - Status: OK

cluster-east: Trial license, expires February 28, 2026 (23 days remaining) - Shards: ¾ used - Status: WARNING - Expiring soon

cluster-central: Trial license, expires February 20, 2026 (15 days remaining) - Shards: ¾ used - Status: WARNING - Expiring soon

You: What's my total memory allocation across all clusters?

AI: Total memory across 3 clusters: 1.1 GB

Cluster Databases Memory
cluster-west 2 200 MB
cluster-east 3 350 MB
cluster-central 3 550 MB

Multi-Cluster Workflow

When configured with multiple profiles, the AI can query across clusters in two ways:

Every Enterprise tool accepts an optional profile parameter. The AI can target any cluster directly without changing defaults:

"Get the license status from cluster-west"
→ AI calls get_license with profile="cluster-west"

"List databases on cluster-east"
→ AI calls list_enterprise_databases with profile="cluster-east"

"Check alerts across all my clusters"
→ AI calls list_alerts for each profile, aggregates results

Profile Switching (Legacy)

Alternatively, the AI can switch the default profile:

  1. List profiles - AI discovers your configured clusters
  2. Set default - AI switches to a cluster using profile_set_default_enterprise
  3. Query - AI runs tools against the current default
  4. Repeat - AI switches and queries other clusters
  5. Aggregate - AI combines results into a unified view

Direct profile targeting is more efficient for cross-cluster queries since it doesn't require switching defaults between each call.

This enables powerful cross-cluster operations that aren't possible with the native Redis Enterprise UI.

Troubleshooting

"Connection refused"

  • Verify the cluster URL and port (default: 9443)
  • Check firewall rules allow access
  • Ensure the REST API is enabled

"Authentication failed"

  • Verify username and password
  • Check the user has appropriate permissions

"Certificate error"

  • Add --insecure for self-signed certificates
  • Or configure a CA certificate with --ca-cert

Tools not appearing

  • Restart your AI assistant after config changes
  • Check that redisctl-mcp is in your PATH

Next Steps