Changelog

Safer retries and clearer request diagnostics.

• Added per-route retry budgets for upstream 502, 503, and 504 responses.
• Added X-Gateway-Attempts to proxy responses.
• Request inspection now separates gateway time from upstream connect and response time.
• Fixed cursor pagination when two services share the same display name.

Regional controls for latency-sensitive workloads.

• Added preferred-region and fallback-region policy fields.
• Introduced the Sydney edge location for Team and Scale workspaces.
• Log exports can now include selected response headers.
• Reduced median policy evaluation time by 7 ms.

More precise rate limits and token scope.

• Added route-specific token scopes for read, write, and proxy operations.
• Rate-limit responses now include RateLimit-Policy.
• Added rolling one-minute quota windows for bursty workloads.
• Fixed idempotency comparisons for equivalent JSON with different key order.

Initial generally available release.

• Stable service, route, proxy, and request inspection endpoints.
• API key authentication and environment-scoped tokens.
• Health-aware routing across North America, Europe, and Asia Pacific.
• Structured request logs with 30-day retention.