[test-improver] Improve tests for launcher/health_monitor

internal/launcher/health_monitor_test.go

@@ -2,6 +2,9 @@ package launcher
 
 import (
 	"context"
+	"encoding/json"
+	"net/http"
+	"net/http/httptest"
 	"testing"
 	"time"
 
@@ -135,3 +138,81 @@ func TestHealthMonitor_RespectsContextCancellation(t *testing.T) {
 		t.Fatal("health monitor did not stop after context cancellation")
 	}
 }
+
+// TestHealthMonitor_MaxFailuresThresholdReached verifies that when a restart failure
+// increments the counter to exactly maxConsecutiveRestartFailures, the counter
+// is capped at the threshold and the server remains in an error state.
+func TestHealthMonitor_MaxFailuresThresholdReached(t *testing.T) {
+	servers := map[string]*config.ServerConfig{
+		"bad-server": {Type: "stdio", Command: "nonexistent-binary-xyz"},
+	}
+	l := newTestLauncher(servers)
+	l.recordError("bad-server", "persistent failure")
+
+	hm := NewHealthMonitor(l, time.Hour)
+	// Set counter to one below the threshold so the next failure reaches it.
+	hm.consecutiveFailures["bad-server"] = maxConsecutiveRestartFailures - 1
+
+	hm.checkAll()
+
+	// Counter must now equal the threshold (not exceed it).
+	require.Equal(t, maxConsecutiveRestartFailures, hm.consecutiveFailures["bad-server"])
+	// The server remains in error state because the restart failed.
+	state := l.GetServerState("bad-server")
+	assert.Equal(t, "error", state.Status)
+}
+
+// mockMCPHandler returns an HTTP handler that responds to any POST with a
+// minimal valid JSON-RPC initialize response. This is sufficient for the
+// plain JSON-RPC transport fallback inside NewHTTPConnection to succeed.
+func mockMCPHandler(t *testing.T) http.Handler {
+	t.Helper()
+	return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		resp := map[string]interface{}{
+			"jsonrpc": "2.0",
+			"id":      1,
+			"result": map[string]interface{}{
+				"protocolVersion": "2024-11-05",
+				"capabilities":    map[string]interface{}{},
+				"serverInfo": map[string]interface{}{
+					"name":    "mock-server",
+					"version": "1.0.0",
+				},
+			},
+		}
+		w.Header().Set("Content-Type", "application/json")
+		data, err := json.Marshal(resp)
+		if err != nil {
+			http.Error(w, "marshal failed", http.StatusInternalServerError)
+			return
+		}
+		_, _ = w.Write(data)
+	})
+}
+
+// TestHealthMonitor_SuccessfulRestartResetCounter verifies that when a
+// health-check restart succeeds, the consecutive-failure counter is reset to
+// zero and the server transitions back to "running" state.
+func TestHealthMonitor_SuccessfulRestartResetCounter(t *testing.T) {
+	mockServer := httptest.NewServer(mockMCPHandler(t))
+	defer mockServer.Close()
+
+	servers := map[string]*config.ServerConfig{
+		"good-server": {Type: "http", URL: mockServer.URL, ConnectTimeout: 1},
+	}
+	l := newTestLauncher(servers)
+
+	// Simulate a server that previously failed.
+	l.recordError("good-server", "transient failure")
+
+	hm := NewHealthMonitor(l, time.Hour)
+	hm.consecutiveFailures["good-server"] = 1
+
+	hm.checkAll()
+
+	// A successful restart must reset the failure counter.
+	assert.Equal(t, 0, hm.consecutiveFailures["good-server"])
+	// The server should now report as running.
+	state := l.GetServerState("good-server")
+	assert.Equal(t, "running", state.Status)
+}