Skip to content

Enhanced ChatGPT Clone: Features Agents, MCP, Skills, DeepSeek, Anthropic, AWS, OpenAI, Responses API, Azure, Groq, o1, GPT-5, Mistral, OpenRouter, Vertex AI, Gemini, Artifacts, AI model switching, message search, Code Interpreter, langchain, DALL-E-3, OpenAPI Actions, Functions, Secure Multi-User Auth, Presets, open-source for self-hosting. Active

⏱️ ci: Give the E2E Redis Transport Lane Its Own Timing Budget (#14907)

Danny Avilae34d83bf7afc
diff --git a/e2e/playwright.config.redis.ts b/e2e/playwright.config.redis.ts
index 548cb262db..e959befc2b 100644
--- a/e2e/playwright.config.redis.ts
+++ b/e2e/playwright.config.redis.ts
@@ -4,6 +4,16 @@ import mockConfig from './playwright.config.mock';
/** Browser scenarios whose behavior crosses the generation stream-store boundary. */
export default defineConfig({
...mockConfig,
+ /**
+ * Every stream event crosses a real Redis round-trip on this lane, so waits that
+ * sit comfortably against the in-memory store run much closer to their budget —
+ * pause/resume and rehydrate paths most of all, since they replay a whole job.
+ * The suite is serial (`workers: 1`), so this is per-operation latency rather
+ * than contention, and it is why this lane reports flaky runs the memory shards
+ * never produce. Give it a larger default assertion budget and one more retry.
+ */
+ retries: process.env.CI ? 3 : 0,
+ expect: { ...mockConfig.expect, timeout: 20_000 },
testMatch: [
/completion\.spec\.ts/,
/deferred-tools-hitl\.spec\.ts/,