4 hours ago
For a service using Railway’s built-in CDN on a Railway-provided domain, does the published approximately 11,000 HTTP requests/second per-domain limit count responses served entirely from cache, including HIT, STALE and cache-generated 304 responses?
Is this enforced globally per domain, per POP, or per edge node? What burst allowance applies, and can the limit be increased for a planned traffic spike?
1 Replies
Status changed to Awaiting Railway Response Railway • about 4 hours ago
3 hours ago
The published approximately 11,000 RPS per-domain figure isn’t enough to determine a safe traffic ceiling for a CDN workload. Cache coverage and requests that reach your service both matter, so we shouldn’t give you a guaranteed cache allowance or burst budget from that figure alone.
What peak RPS and duration are you expecting, when is the spike, and where will most visitors be located? If you have an estimated cache-hit ratio, include that too. Those details will help us assess the capacity you need and whether a limit adjustment is appropriate; an increase isn’t guaranteed.
Status changed to Awaiting User Response Railway • about 3 hours ago