Unverified50% confidenceDecision RuleExact time
Setting maxRetries to 0 in generationConfig removes SDK-layer retries, so an endpoint that keeps returning 503 receives only 7 requests (about 74 seconds) before ending with [API Error: 503 upstream busy] and exit code 1.
1
Sources
50%
Confidence
Long-term
Relevance
10/18/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedAfter setting DFLASH_TOKENS=15, in citation requests with a 20k context it causes the engine to crash, reporting EngineDeadError and a marlin_gemm allocation failure, after which every request returns 500 and it cannot self-recover.62% similarUnverifiedv0.10.0 introduced login rate limiting: 10 wrong API Keys within 5 minutes trigger a 15-minute 429 lockout and return a Retry-After header, support proxy awareness via TRUSTED_PROXIES, and can be disabled with AUTH_THROTTLE=off.61% similarUnverifiedv3.23.0 fixes issues in the responses module related to avoiding replay of indeterminate typed sends on reconnect (#4011) and respecting send queue limits during reconnection (#4000).61% similarUnverifiedRetries for free APIs should only target transient errors such as 429 and 503; other errors should be thrown directly to avoid blindly retrying code bugs.58% similarUnverifiedThe SDK runtime source code of v0.23.1 differs from 0.23.0 only in the version number string, with API snapshots, dependencies, supported Python versions, and persistence schema all remaining unchanged.58% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1064325API
curl https://kongchang.com/api/v1/knowledge/claims/1064325MCP
get_claim(id=1064325)