Files
coder/coderd/mcp_http.go
Bobby Ho 603dd420d1 feat: add ServerTailnet logs and metrics for unreachable agents (#30133)
Add default-level metrics and warn-level diagnostics to investigate
ServerTailnet reachability failures causing agent-specific 502s, without
restarting coderd or enabling debug logging.

- Enrich request logs and PTY/chatd warnings with agent ID, failure
  reason, wait duration, coordinator state, WireGuard handshake and
  byte counts, and DERP regions.
- Add metrics for unreachable attempts by reason, reachability latency,
  embedded DERP connections, and offline app/terminal requests.
- Distinguish reachability failures from subsequent TCP dial failures,
  cap reverse proxy dials at one minute, and exclude expected chatd
  readiness polls and superseded-agent dials from unreachable counts.
- Report missing handshakes as zero time instead of 1970, also fixing
  `coder ping`.

Part of PLAT-680, a sub-issue of PLAT-360.
2026-10-14 14:55:29 -07:00

115 lines
4.4 KiB
Go

package coderd
import (
"context"
"errors"
"fmt"
"net/http"
"github.com/google/uuid"
"golang.org/x/xerrors"
"cdr.dev/slog/v3"
"github.com/coder/coder/v2/coderd/database/dbauthz"
"github.com/coder/coder/v2/coderd/httpapi"
"github.com/coder/coder/v2/coderd/httpmw"
"github.com/coder/coder/v2/coderd/mcp"
"github.com/coder/coder/v2/coderd/rbac/policy"
"github.com/coder/coder/v2/coderd/workspaceapps"
"github.com/coder/coder/v2/codersdk"
"github.com/coder/coder/v2/codersdk/toolsdk"
"github.com/coder/coder/v2/codersdk/workspacesdk"
)
type MCPToolset string
const (
MCPToolsetStandard MCPToolset = "standard"
MCPToolsetChatGPT MCPToolset = "chatgpt"
)
// mcpHTTPHandler creates the MCP HTTP transport handler
// It supports a "toolset" query parameter to select the set of tools to register.
func (api *API) mcpHTTPHandler() http.Handler {
return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
// Create MCP server instance for each request
mcpServer, err := mcp.NewServer(api.Logger.Named("mcp"))
if err != nil {
api.Logger.Error(r.Context(), "failed to create MCP server", slog.Error(err))
httpapi.Write(r.Context(), w, http.StatusInternalServerError, codersdk.Response{
Message: "MCP server initialization failed",
})
return
}
// Keep tool requests in-process. The private transport carries the
// originating MCP audience without granting external REST access.
transport, closeTransport := newMCPDelegatedTransport(api.Logger, api.RootHandler, api.AccessURL, httpmw.APIKey(r).ID)
defer closeTransport()
// The MCP SDK detaches tool cancellation and waits for in-flight tools
// before returning, so close their transport independently on disconnect.
stopClose := context.AfterFunc(r.Context(), closeTransport)
defer stopClose()
authenticatedClient := codersdk.New(api.AccessURL,
codersdk.WithSessionToken(httpmw.APITokenFromRequest(r)),
codersdk.WithHTTPClient(&http.Client{Transport: transport}))
// Wrap the agent connection function to enforce ActionSSH
// on the workspace. Without this check, a user who can read
// a workspace but lacks SSH permission could still execute
// commands through MCP tools.
toolOpt := toolsdk.WithAgentConnFunc(func(ctx context.Context, agentID uuid.UUID) (workspacesdk.AgentConn, func(), error) {
if api.Entitlements.Enabled(codersdk.FeatureBrowserOnly) {
return nil, nil, xerrors.New("non-browser connections are disabled")
}
// Use system context for the lookup because the tool
// handler context does not carry a dbauthz actor. The
// real authorization happens in the Authorize call below.
//nolint:gocritic // The system query only fetches the workspace
// object so we can perform an ActionSSH check against it
// with the real user's roles via api.Authorize.
workspace, err := api.Database.GetWorkspaceByAgentID(dbauthz.AsSystemRestricted(ctx), agentID)
if err != nil {
return nil, nil, xerrors.Errorf("get workspace by agent ID: %w", err)
}
// Enforce the same ActionSSH check that the coordinate
// endpoint uses (workspaceagents.go:1317).
if !api.Authorize(r, policy.ActionSSH, workspace) {
return nil, nil, xerrors.New("unauthorized: you do not have SSH access to this workspace")
}
conn, release, err := api.agentProvider.AgentConn(ctx, agentID)
if unreachable, ok := errors.AsType[*workspaceapps.AgentUnreachableError](err); ok {
// Tool errors are sent with HTTP 200, so the request log
// line is written at debug level.
api.Logger.Warn(ctx, "agent is unreachable", append(unreachable.Fields, slog.Error(err))...)
}
return conn, release, err
})
toolset := MCPToolset(r.URL.Query().Get("toolset"))
// Default to standard toolset if no toolset is specified.
if toolset == "" {
toolset = MCPToolsetStandard
}
switch toolset {
case MCPToolsetStandard:
if err := mcpServer.RegisterTools(authenticatedClient, toolOpt); err != nil {
api.Logger.Warn(r.Context(), "failed to register MCP tools", slog.Error(err))
}
mcpServer.RegisterPrompts()
case MCPToolsetChatGPT:
if err := mcpServer.RegisterChatGPTTools(authenticatedClient, toolOpt); err != nil {
api.Logger.Warn(r.Context(), "failed to register MCP tools", slog.Error(err))
}
default:
httpapi.Write(r.Context(), w, http.StatusBadRequest, codersdk.Response{
Message: fmt.Sprintf("Invalid toolset: %s", toolset),
})
return
}
// Handle the MCP request
mcpServer.ServeHTTP(w, r)
})
}