A development team experienced a 26-minute outage when DeepSeek retired two API model names, deepseek-chat and deepseek-reasoner, replacing them with deepseek-v4-pro and deepseek-v4-flash. The team's LLM gateway continued to use the old model names, resulting in HTTP 400 errors for all model calls. Their configured fallback mechanism failed because it used the same provider and API key, meaning it was also affected by the model name change. The issue was resolved by updating the gateway configuration to use the new model names, and the team identified a design flaw in their fallback strategy, noting that true resilience requires crossing provider boundaries. AI
IMPACT Highlights the fragility of LLM integrations and the need for robust fallback strategies that cross provider boundaries.
RANK_REASON The item describes a specific technical issue and its resolution related to an LLM provider's API changes, impacting a user's system.
- 400 Bad Request
- continuous integration
- DeepSeek
- deepseek-chat
- deepseek-reasoner
- deepseek-v4-flash
- deepseek-v4-pro
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
