{"version":"2.1.309","anchor":"patient-tide-retry-budget-gate-for-main-thread-requests","canonical_anchor":"patient-tide-retry-budget-gate-for-main-thread-requests","heading":"Server-controlled extra retries for requests in the main conversation","tier":"soon","area":"API Requests","scope":"individual","heads_up":false,"url":"https:\/\/changelogs.core-directive.com\/v\/2.1.309\/e\/patient-tide-retry-budget-gate-for-main-thread-requests","release_url":"https:\/\/changelogs.core-directive.com\/v\/2.1.309","markdown":"### Server-controlled extra retries for requests in the main conversation\n\nClaude Code can now get an extra retry allowance from the server for main-conversation requests, unless you set a retry variable\n\n**What**\n\nWhen a request to the model fails, Claude Code can try it again. There is now a separate retry allowance whose size the server sets. It is zero, and so has no effect, unless the server sends a value. It only applies to requests from the main conversation, not to background work. Claude Code accepts a positive whole number and caps it at a limit.\n\nThe allowance is always zero when either of these environment variables is set:\n\n- `CLAUDE_CODE_MAX_RETRIES`\n\n- `CLAUDE_CODE_RETRY_WATCHDOG`\n\n**Why**\n\nThis lets the server give main-conversation requests a few more tries. Setting your own retry variable switches it off, so your setting is the one that counts.\n\n- Flag `tengu_patient_tide`: Not enough to say (read for one account on one subscription tier against v2.1.309; this account: no value returned, anonymous baseline: no value returned, compiled default: not a boolean we can read) These values were read against a different version of Claude Code, so treat them as the nearest reading available instead of one taken on this release.\n- Area: API Requests\n- Tier: Nothing to try yet\n- Useful: 2\/5\n- Signal: 3\/5\n- Scope: individual\n- Heads-up: no\n- Present in the build but not switched on"}