Analytics
Every tool call through the workspace gateway is logged to Cloudflare Analytics Engine. You get full visibility into what tools are being called, by whom, how often, and how fast.
What Gets Tracked
Section titled “What Gets Tracked”Every tools/call request records:
| Field | Description |
|---|---|
token_id | Which token made the call (identifies the member or integration) |
workspace_id | Which workspace the call was routed through |
server_slug | Which MCP server handled the call |
tool_name | The specific tool that was called |
duration_ms | How long the upstream server took to respond |
status | Success or failure |
timestamp | When the call occurred |
Viewing Analytics
Section titled “Viewing Analytics”Dashboard
Section titled “Dashboard”Open your workspace in the Admin dashboard and click the Analytics tab. The dashboard shows:
Overview Panel
- Total tool calls (today, this week, this month)
- Success rate
- Average response time
- Active tokens
Tool Breakdown
- Calls per tool, sorted by frequency
- Average duration per tool
- Error rate per tool
Member Activity
- Calls per member (identified by token name)
- Most-used tools per member
- Activity timeline
Server Health
- Calls per server
- Average latency per server
- Error rate per server
Use Cases
Section titled “Use Cases”Identify Popular Tools
Section titled “Identify Popular Tools”See which tools your team uses most. This helps you decide which MCP servers are worth keeping, which need better configuration, and where to invest in custom tools.
Debug Slow Tools
Section titled “Debug Slow Tools”The duration breakdown shows which tools are slow. If a server consistently takes more than a few seconds, you may need to optimize it, increase its timeout, or replace it with a faster alternative.
Monitor Bot Behavior
Section titled “Monitor Bot Behavior”If a workspace powers a bot, analytics show exactly how the bot is using tools. You can see if the bot is calling the right tools, if it is making excessive calls, or if certain tools are failing.
Track Team Usage
Section titled “Track Team Usage”With per-member analytics, you can see who is using the workspace, how often, and which tools they rely on. This is useful for understanding adoption and identifying team members who might need help getting started.
Billing Visibility
Section titled “Billing Visibility”Tool calls count toward your plan’s MCP proxy call limit. The analytics dashboard shows your current usage against your limit, so you can anticipate when you might need to upgrade.
Analytics Retention
Section titled “Analytics Retention”| Plan | Retention |
|---|---|
| Free | 7 days |
| Starter | 30 days |
| Pro | 90 days |
| Business | 1 year |
| Enterprise | Custom |
Querying Analytics Programmatically
Section titled “Querying Analytics Programmatically”Analytics data is available through the Aerostack API for programmatic access:
# Get workspace analytics summarycurl https://api.aerostack.dev/api/community/mcp/workspaces/WORKSPACE_ID/analytics \ -H "Authorization: Bearer YOUR_JWT"The response includes aggregated metrics for the requested time range.
Observability in the Dashboard
Section titled “Observability in the Dashboard”Beyond analytics, the dashboard provides a Logs view for real-time debugging:
- Live log stream — see tool calls as they happen
- Error details — view the full error message and upstream response for failed calls
- Latency waterfall — understand where time is spent (gateway processing vs. upstream server)
- Filter by tool, server, or member — narrow down to specific calls
This is particularly useful when setting up a new workspace or debugging why a tool is not working as expected.