Configure PromptKit transient retries

internal/framework/llm/promptkit_client.go

@@ -26,15 +26,16 @@ type PromptKitLocalBackendConfig struct {
 }
 
 type PromptKitClientConfig struct {
-	ProfileDir      string
-	ProfileFile     string
-	LocalBackend    *PromptKitLocalBackendConfig
-	Assets          *AssetRegistry
-	Timeout         time.Duration
-	HTTPClient      *http.Client
-	EngineOptions   []promptkit.Option
-	Recorder        *LLMProfileRecorder
-	ReasoningEffort *string
+	ProfileDir       string
+	ProfileFile      string
+	TransientRetries *int
+	LocalBackend     *PromptKitLocalBackendConfig
+	Assets           *AssetRegistry
+	Timeout          time.Duration
+	HTTPClient       *http.Client
+	EngineOptions    []promptkit.Option
+	Recorder         *LLMProfileRecorder
+	ReasoningEffort  *string
 }
 
 type PromptKitClient struct {
@@ -79,9 +80,10 @@ func NewPromptKitClient(cfg PromptKitClientConfig) (*PromptKitClient, error) {
 	options = append(options, cfg.EngineOptions...)
 
 	engine, err := promptkit.NewEngine(promptkit.Config{
-		ProfileDir: profileSource.ProfileDir,
-		Timeout:    cfg.Timeout,
-		HTTPClient: cfg.HTTPClient,
+		ProfileDir:       profileSource.ProfileDir,
+		Timeout:          cfg.Timeout,
+		HTTPClient:       cfg.HTTPClient,
+		TransientRetries: cfg.TransientRetries,
 	}, options...)
 	if err != nil {
 		return nil, fmt.Errorf("create PromptKit engine: %w", err)

internal/framework/llm/promptkit_client_test.go

@@ -971,6 +971,70 @@ model: local-model
 	}
 }
 
+func TestPromptKitClientAppliesConfiguredBuiltInTransientRetries(t *testing.T) {
+	for _, test := range []struct {
+		name      string
+		retries   int
+		statuses  []int
+		wantCalls int32
+		wantErr   bool
+	}{
+		{name: "eligible failure then success", retries: 1, statuses: []int{http.StatusServiceUnavailable, http.StatusOK}, wantCalls: 2},
+		{name: "disabled", retries: 0, statuses: []int{http.StatusServiceUnavailable}, wantCalls: 1, wantErr: true},
+		{name: "exhausted", retries: 1, statuses: []int{http.StatusServiceUnavailable, http.StatusServiceUnavailable}, wantCalls: 2, wantErr: true},
+		{name: "noneligible client error", retries: 2, statuses: []int{http.StatusBadRequest}, wantCalls: 1, wantErr: true},
+	} {
+		t.Run(test.name, func(t *testing.T) {
+			var calls atomic.Int32
+			server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {
+				index := calls.Add(1) - 1
+				if int(index) >= len(test.statuses) {
+					t.Errorf("unexpected HTTP attempt %d", index+1)
+					w.WriteHeader(http.StatusInternalServerError)
+					return
+				}
+				status := test.statuses[index]
+				w.Header().Set("Content-Type", "application/json")
+				if status != http.StatusOK {
+					w.WriteHeader(status)
+					_, _ = w.Write([]byte(`{"error":{"message":"provider failure"}}`))
+					return
+				}
+				_, _ = w.Write([]byte(`{"choices":[{"message":{"content":"{\"ok\":true}"}}]}`))
+			}))
+			defer server.Close()
+
+			profilePath := filepath.Join(t.TempDir(), "profile.yml")
+			if err := os.WriteFile(profilePath, []byte("id: retry-profile\nendpoint: "+server.URL+"/v1\nmodel: retry-model\n"), 0o600); err != nil {
+				t.Fatal(err)
+			}
+			client, err := NewPromptKitClient(PromptKitClientConfig{
+				Assets:           newTestPromptKitAssets(t),
+				ProfileFile:      profilePath,
+				TransientRetries: &test.retries,
+			})
+			if err != nil {
+				t.Fatal(err)
+			}
+			var out map[string]any
+			_, err = client.CompleteStructured(context.Background(), contracts.StructuredCompletionRequest{
+				PromptID:  "adapter.test",
+				ProfileID: "retry-profile",
+				SessionID: "retry-test",
+				Inputs: contracts.LLMInputSet{
+					"transcript": contracts.NewLLMInputMaterial("transcript", "application/json", []byte(`{"source":true}`), "", ""),
+				},
+			}, &out)
+			if (err != nil) != test.wantErr {
+				t.Fatalf("CompleteStructured() error = %v, want error=%t", err, test.wantErr)
+			}
+			if got := calls.Load(); got != test.wantCalls {
+				t.Fatalf("HTTP calls = %d, want %d", got, test.wantCalls)
+			}
+		})
+	}
+}
+
 func TestLLMProfileRecorderDistinguishesEffectiveTargets(t *testing.T) {
 	recorder := NewLLMProfileRecorder()
 	for _, profile := range []artifacts.LLMProfileManifest{