Fix metric naming inconsistencies and typos

- Remove _total suffix from counter names (auto-added by Prometheus exporter)
- Add metric.WithUnit("s") for histogram durations instead of _seconds suffix
- Fix spelling: "exection" → "execution" in docs
- Update middleware chain comment to include telemetry

Co-authored-by: jerm-dro <10532181+jerm-dro@users.noreply.github.com>

Author: Copilot

docs/operator/virtualmcpserver-observability.md

@@ -39,7 +39,7 @@ the workflow name as an attribute.
 
 ## Distributed Tracing
 
-The vMCP creates spans for each individual backend operation as well as workflow executions, enabling the attribution of workflow exection errors or latency to specific tool calls.
+The vMCP creates spans for each individual backend operation as well as workflow executions, enabling the attribution of workflow execution errors or latency to specific tool calls.
 
 
 ## Configuration

pkg/vmcp/server/server.go

@@ -385,7 +385,7 @@ func (s *Server) Start(ctx context.Context) error {
 		logger.Info("RFC 9728 OAuth discovery endpoints enabled at /.well-known/")
 	}
 
-	// MCP endpoint - apply middleware chain: auth → discovery
+	// MCP endpoint - apply middleware chain: auth → discovery → telemetry
 	var mcpHandler http.Handler = streamableServer
 
 	if s.config.TelemetryProvider != nil {

pkg/vmcp/server/telemetry.go

@@ -39,22 +39,21 @@ func monitorBackends(
 	backendCount.Record(ctx, int64(len(backends)))
 
 	requestsTotal, err := meter.Int64Counter(
-		"toolhive_vmcp_backend_requests_total",
-		metric.WithDescription("Total number of requests per backend"),
-	)
+		"toolhive_vmcp_backend_requests",
+		metric.WithDescription("Total number of requests per backend"))
 	if err != nil {
 		return nil, fmt.Errorf("failed to create requests total counter: %w", err)
 	}
 	errorsTotal, err := meter.Int64Counter(
-		"toolhive_vmcp_backend_errors_total",
-		metric.WithDescription("Total number of errors per backend"),
-	)
+		"toolhive_vmcp_backend_errors",
+		metric.WithDescription("Total number of errors per backend"))
 	if err != nil {
 		return nil, fmt.Errorf("failed to create errors total counter: %w", err)
 	}
 	requestsDuration, err := meter.Float64Histogram(
 		"toolhive_vmcp_backend_requests_duration",
 		metric.WithDescription("Duration of requests in seconds per backend"),
+		metric.WithUnit("s"),
 	)
 	if err != nil {
 		return nil, fmt.Errorf("failed to create requests duration histogram: %w", err)
@@ -161,24 +160,25 @@ func monitorWorkflowExecutors(
 	meter := meterProvider.Meter(instrumentationName)
 
 	executionsTotal, err := meter.Int64Counter(
-		"toolhive_vmcp_workflow_executions_total",
+		"toolhive_vmcp_workflow_executions",
 		metric.WithDescription("Total number of workflow executions"),
 	)
 	if err != nil {
 		return nil, fmt.Errorf("failed to create workflow executions counter: %w", err)
 	}
 
 	errorsTotal, err := meter.Int64Counter(
-		"toolhive_vmcp_workflow_errors_total",
+		"toolhive_vmcp_workflow_errors",
 		metric.WithDescription("Total number of workflow execution errors"),
 	)
 	if err != nil {
 		return nil, fmt.Errorf("failed to create workflow errors counter: %w", err)
 	}
 
 	executionDuration, err := meter.Float64Histogram(
-		"toolhive_vmcp_workflow_duration_seconds",
+		"toolhive_vmcp_workflow_duration",
 		metric.WithDescription("Duration of workflow executions in seconds"),
+		metric.WithUnit("s"),
 	)
 	if err != nil {
 		return nil, fmt.Errorf("failed to create workflow duration histogram: %w", err)