Edge delivery infrastructure
Flyner operates a compact edge network for API and media workloads. We keep routing simple, capacity honest and the operational surface small — so the numbers you see in staging hold up in production.
We do four things well and integrate with the rest of your stack instead of replacing it.
Requests collapse at the edge, so your origin only sees the traffic that matters. Cache rules are versioned alongside your deploy.
HTTP/2 and HTTP/3 with per-route timeouts, buffering disabled where streaming is expected, and no hidden middleboxes.
Traffic enters through the nearest point of presence and leaves through the region you choose, with the path documented end to end.
Structured access logs, request IDs and percentiles we actually publish. No dashboards that disagree with each other.
Edge nodes terminate TLS and hand requests to the origin over a persistent connection. Health checks are separate from customer routes and never count against your quota.
Every deployment exposes /health for monitors and load balancers. It returns a static JSON document and never touches the origin.
Routes carrying long-lived streams run with buffering off and a 600 second read timeout. Chunked uploads pass through untouched.
Configuration changes roll out per region. Rollbacks are a single command and complete in under two minutes.