Observe
MCP & Tools
The MCP & Tools page shows how MCP servers and tools are used across the organization. Use it to identify which servers and tools deliver value, spot where failures cluster, and watch adoption per source and skill. Open it from Observe > MCP & Tools in the dashboard, and switch between the Tools and MCP Servers sub-tabs.
Access requirements
Section titled “Access requirements”Viewing this page requires the org:admin scope. Access is included in the default Admin role but not the Member role. For Member users, the nav entry stays visible, but opening the page shows an access-restricted notice.
Headline metrics
Section titled “Headline metrics”The top of the page summarizes the selected period:
- Avg Success Rate shows the share of tool calls that completed successfully
- Total Events counts tool call volume
- Active Users counts members generating tool activity
- Active Targets counts MCP servers and tool sources in use
- Unique Tools counts distinct tools called
Which servers and tools deliver value
Section titled “Which servers and tools deliver value”The Activity by Source and Users by Source charts show which MCP servers and tool sources drive volume and reach. A source with high event counts but few users serves a narrow audience; a source with broad user coverage is delivering value across the team. User Usage and User Event Counts break activity down per user, and Skill Usage and Users per Skill show how each skill is adopted and by how many people.
Clicking a bar segment or row drills down. The clicked server or user becomes an active filter chip, so the view narrows from an aggregate chart to a single source or user without rebuilding the query. Time-series charts also support drag-zoom, which re-scopes the page’s date range to the selected region.
Spotting and investigating failures
Section titled “Spotting and investigating failures”Two charts surface where things break:
- Errors Over Time plots failure counts over time, and its tooltip lists the top failing tools at each point
- Failures by Source and Tool aggregates failures per server and tool, showing only rows with recorded failures
To investigate a failure cluster, start with Errors Over Time to find when failures spiked, move to Failures by Source and Tool to see which server and tool are responsible, click through to filter the page to that source, then open the pre-filtered Tool Logs view to inspect individual calls. Failure counts include calls blocked by policy as well as runtime errors, so a spike can indicate policy enforcement at work rather than a broken tool.
Filters
Section titled “Filters”The filter bar narrows every chart on the page. Filter by server, user email, agent, role, account type, status, and date range. The Type filter distinguishes Hosted MCP Servers, Tunneled MCP Servers, Shadow MCP Servers, Local Tools, and Skills, which is useful for watching adoption of sanctioned servers versus shadow or local usage. The AI-assistant mode answers questions about server and tool usage in natural language.
Empty state
Section titled “Empty state”Before any data arrives, the page shows an Install Observability plugin prompt with setup flows for Claude, Cursor, and Codex. Once agents report activity, the dashboards populate automatically. If existing filters match nothing, the page suggests adjusting the search query or time range.