fix(similarasin): LLM 切换 gemini-3.8-flash + 修复双实例 stale 判死盲区

- 模型切换:glm-5.3-flash / gemini-3.7-flash(类目/合规/图片对比,appearancepatent)→ gemini-3.8-flash
  涉及 SimilarAsinProperties、AppearancePatentProperties、application.yml、TLS 探针与测试断言
- stale 判死盲区:finalizeStaleTasks 去掉 isOwnerCurrent 过滤,改由 job 锁持有实例全局判死
  (任务 27117 线上死因:客户端 done 信号被任务锁拒后断线,owner 过滤导致无人判死,
  被另一实例 job stuck 扫描误判为"文件生成任务运行超时";appearancepatent 同源盲区一并修复)
- 客户端回传重试:3x2s → 8 次,40902 锁忙固定等 15s(app_client/similar_asin.py,随客户端发包生效)
This commit is contained in:
2026-09-07 22:10:20 +08:00
parent d9d2c1692e
commit 499841dfa5
8 changed files with 41 additions and 19 deletions
@@ -61,7 +61,7 @@ public class LlmGatewayTlsProbe {
}
HttpClient client = builder.build();
try {
String body = "{\"model\":\"glm-5.3-flash\",\"messages\":[{\"role\":\"user\",\"content\":\"ping\"}],\"max_tokens\":10}";
String body = "{\"model\":\"gemini-3.8-flash\",\"messages\":[{\"role\":\"user\",\"content\":\"ping\"}],\"max_tokens\":10}";
HttpRequest request = HttpRequest.newBuilder()
.uri(URI.create("https://" + host + "/v1/chat/completions"))
.timeout(Duration.ofSeconds(30))