Skip to main content
Once your gateways are live, the Logs section gives you complete visibility into every tool call flowing through them. Click any log entry to open its full trace — request, response, headers, and the gateway/instance/user context that produced it.

​
The Logs screen

Open Logs from the left sidebar. You’ll see a reverse-chronological table of every tool call across your gateways, with these columns:
  • Session ID — the proxy session the call belongs to (e.g., proxy-20b3c2e6-e24b-4183-…)
  • Timestamp — when the call was made
  • Server Name / Entity Name — the MCP server logo (Hubspot, Notion, Linear, etc.) and the tool that was invoked (e.g., HUBSPOT_LIST_CONTACTS, get_me, search)
  • Method — the MCP method used (e.g., tools/call)
  • Gateway Name — which Gateway Config the instance belongs to (e.g., “Daily work”, “Notion”)
  • Instance Name — the specific instance that handled the call
  • Client Name — the MCP client that connected (e.g., demo-agent, Claude Desktop, Cursor)
Logs Screen

​
Filtering logs

The top bar gives you two main ways to narrow results:
  • Filters — a count badge shows how many filters are active; click to add filters on server, tool, gateway, instance, client, or session. Use Clear all to reset.
  • Date & Time — defaults to All time; switch to a specific window (last hour, last 24 hours, custom range) to focus on recent activity.
Use the refresh icon on the right to pull the latest entries without reloading the page.

​
Opening a trace

Click any log row to open the Trace details panel on the right. Every trace starts with an Info block showing:
  • Session ID
  • Entity Name (the tool invoked)
  • Duration (e.g., 211.0 ms)
  • Timestamp
  • Method (e.g., tools/call)
  • Client Name
  • Client Version
Below that, a set of collapsible sections gives you the full call context:
  • Request (JSON) — the exact input the MCP client sent
  • Response (JSON) — what the gateway returned to the client
  • Headers (JSON) — the MCP headers accompanying the call
  • MCP Gateway — which Gateway Config handled the call
  • Instance — which Instance served it
  • User Details — the connected user context
This is the view you’ll use to debug a broken tool call, confirm the right account was used, or audit exactly what an MCP client asked for and received.
Trace Details 1

​
Common use cases

  • Debugging — when a client gets an unexpected result, open the trace and inspect the request/response JSON to pinpoint what went wrong.
  • Auditing — produce a full history of which clients invoked which tools on which gateways, with millisecond-level timestamps.
  • Performance tuning — scan durations across similar calls to identify slow tools or upstream APIs.
  • Security review — filter by client or server to confirm that activity matches expectations.

​
Retention

Logs and traces are retained according to your workspace’s plan. For long-term storage or compliance archives, export trace data out of Synqed.