Edge Overview

FoundryDB Edge puts a managed HTTP gateway in front of your application. It is served from European points of presence (Stockholm, Helsinki, and Frankfurt), so requests are handled close to your users while traffic is forwarded to your application origin.

What Edge Provides

Capability	What it does
Custom domains	Bring your own domain and route it through the gateway.
Automatic TLS	Certificates are issued and renewed for your domain automatically.
Edge caching	Cacheable responses are served from the point of presence.
Rate limiting	Protect your origin by capping request rates per client.
Web application firewall (WAF)	Detection of common malicious request patterns.
Request analytics	Per-request visibility into traffic, status codes, and cache behavior.

Resilience and Scale

PoP failover & autoscaling

Nearest PoP serves each client · HA pair + autoscaling

Clientnearest PoPserving address →PoP pairactive + standby⇢ forwardApp originyour app

Active · serving addressHot standbyFailed nodeAdded capacityClient requestForward to origin (dashed)

Beyond the gateway features above, Edge is built to stay reachable and to keep pace with traffic:

• Each point of presence runs as a highly available pair, so a single node failure does not take your domain offline.
• If an entire point of presence becomes unavailable, traffic is automatically repointed to another location.
• Each point of presence scales out as request volume rises and scales back in as it subsides.

See Resilience and Autoscaling for details on how these behaviors work and what you can observe.

Where Edge Runs

Points of presence are located in the EU:

Location	City	UpCloud zone
Stockholm	Sweden	se-sto1
Helsinki	Finland	fi-hel1
Frankfurt	Germany	de-fra1

Each PoP consists of an active node and a standby node in the same zone. Incoming requests are routed to the geographically nearest PoP. The PoP terminates TLS, applies caching and rate-limiting rules, and then forwards the request to your application origin over a private network path. If the active node in a PoP fails, the standby takes over and the serving floating IP moves to it automatically. If an entire PoP becomes unreachable, DNS is updated to direct traffic to another PoP in the fleet.