◉CLAUDE LABJP
●SONNET5 — Claude Sonnet 5 is now the default model in Claude Code, with a native 1M-token context and promo pricing through Aug 31●MCPTUNNEL — MCP tunnels arrives as a Research Preview, letting you reach MCP servers inside private networks●IDP — Admins can now provision MCP connectors via their IdP (starting with Okta); Asana, Figma, Linear and more support managed auth●SANDBOX — Claude Managed Agents can now run tool execution in your own self-hosted sandbox instead of Anthropic's infrastructure●FAST — Fast mode for Opus 4.7 is deprecated and will be removed on July 24; migrate to fast mode for Opus 4.8●FIXES — Recent fixes cover public gateway endpoints, a confirm prompt for external git worktrees, and MCP request timeouts●SONNET5 — Claude Sonnet 5 is now the default model in Claude Code, with a native 1M-token context and promo pricing through Aug 31●MCPTUNNEL — MCP tunnels arrives as a Research Preview, letting you reach MCP servers inside private networks●IDP — Admins can now provision MCP connectors via their IdP (starting with Okta); Asana, Figma, Linear and more support managed auth●SANDBOX — Claude Managed Agents can now run tool execution in your own self-hosted sandbox instead of Anthropic's infrastructure●FAST — Fast mode for Opus 4.7 is deprecated and will be removed on July 24; migrate to fast mode for Opus 4.8●FIXES — Recent fixes cover public gateway endpoints, a confirm prompt for external git worktrees, and MCP request timeouts
TAG

cache-stampede

1 articles
← Back to all tags
Related:
claude-api1singleflight1coalescing1rate-limit1production1cost-optimization1
⬡ API & SDK/2026-07-13Advanced

Coalescing Concurrent Claude API Calls: Single-Flight Against Duplicate Inference and Cache Stampede

A design for collapsing identical prompts that fire at the same instant into a single upstream Claude call, using single-flight (request coalescing). In-process and distributed implementations, jittered retries, and negative caching, with measured results.