The GLMT retry logic tests use dynamic imports and GlmtProxy instantiation which are slower on CI runners than locally. Increase default timeout from 5s to 30s to prevent flaky failures.
Add retry logic and connection pooling to GLMT proxy for handling Z.AI 429 rate limit errors gracefully. Changes: - Add RetryConfig with env vars (GLMT_MAX_RETRIES, GLMT_RETRY_BASE_DELAY, GLMT_DISABLE_RETRY) - Add exponential backoff with jitter and Retry-After header support - Add forwardWithRetry() and forwardAndStreamWithRetry() wrappers - Add HTTPS connection pooling via shared https.Agent - Add comprehensive unit tests (19 tests covering all scenarios) Fixes: #402