supplier+deploy/mainnet: log job_failed reasons; bump openai timeout 180s→300s; chatmock reasoning_effort=low

- jobRunner.ts: console.warn at each jobs.fail site (chat + tts paths, 6 sites). Three mainnet GPT-supplier 502s on 2026-05-19 left no trace beyond Traefik — adding a one-line log restores grep-ability.
- docker-compose.supplier-gpt.yml: OPENAI_TIMEOUT_MS 180000→300000. The 3 observed 502s were all exactly 90 polls × 2s, i.e. the supplier's openai_timeout firing.
- docker-compose.chatmock.yml: serve --reasoning-effort low --reasoning-summary none. Default 'medium'/'auto' pushed GPT-5.4 think-time over the prior 180s budget.

Note: on-chain advert ADVERT_MAX_PROCESSING_MS still pins 180000 until a re-advert tx is posted — buyers compare against the on-chain value, not the runtime env.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

deploy/mainnet/docker-compose.chatmock.yml

@@ -45,7 +45,12 @@ services:
   chatmock:
     image: storagetime/chatmock:v1.36
     container_name: marketplace-mainnet-chatmock
-    command: ["serve"]
+    # --reasoning-effort low + --reasoning-summary none cap GPT-5 thinking
+    # tokens so a normal request returns well inside the supplier's
+    # OPENAI_TIMEOUT_MS budget. Default `medium`/`auto` produced 3-min
+    # timeouts in early mainnet testing (jobs bcc973bd / c85eaeec /
+    # 7e87afe6 on 2026-05-19).
+    command: ["serve", "--reasoning-effort", "low", "--reasoning-summary", "none"]
     restart: unless-stopped
     environment:
       CHATGPT_LOCAL_HOME: /data

deploy/mainnet/docker-compose.supplier-gpt.yml

@@ -36,7 +36,7 @@ services:
     environment:
       LLM_BACKEND: openai
       OPENAI_BASE_URL: http://marketplace-mainnet-chatmock:8000
-      OPENAI_TIMEOUT_MS: "180000"
+      OPENAI_TIMEOUT_MS: "300000"
       PORT: "8080"
       # Enable LiveOgmiosProvider so submitTx/awaitTx actually broadcast.
       # Mirrors the testnet supplier compose; without this the supplier