Optimise performance and availability with intelligent traffic management.
One server is a single point of failure. Our managed load balancer spreads application traffic across multiple origins, terminates SSL, probes every backend continuously and fails over in under a second, so a traffic spike or a dead node never becomes an outage. Layer 4 and Layer 7, on any origin you own.
Unhealthy origin
Out of rotation <1s
Most platforms end up using more than one. A single balancer handles all three, configured per service.
Steer visitors to the closest healthy region automatically. Traffic is routed on live latency and origin health rather than a static DNS record, so a regional outage reroutes in seconds instead of waiting out a TTL.
Where our regions areLayer 7 balancing for HTTP, HTTPS and HTTP/2 with SSL/TLS termination, keep-alive reuse and compression handled at the balancer. Your origins stop spending CPU on handshakes and start spending it on requests.
Pair it with edge cachingRaw TCP and UDP balancing for everything that is not a web app, game servers, mail, databases, message queues, custom protocols. Full transparency, connection multiplexing and high throughput.
See VPS origins to balanceA load balancer is only useful if you trust what it does when something breaks. Here is the whole path, including the part that matters most, the failure case.
Your domain points at the load balancer instead of a single server, so no visitor ever holds the IP address of an individual origin.
TLS is terminated at the balancer. For Layer 7 services the request line, host and headers are read so routing decisions can be made on real content, not just an IP tuple.
Round-robin, least-connections, weighted or IP-hash, whichever fits. Sticky sessions keep a returning visitor on the origin that already holds their session state.
Every origin is probed on an interval. A failing origin is pulled from rotation before a real visitor ever lands on it.
When an origin misses its checks, traffic shifts to the healthy pool in under a second. When it recovers, it is eased back in with slow-start rather than being hit at full load.
Both are included on every plan. This is the practical difference.
| Layer 4 (TCP/UDP) | Layer 7 (HTTP/S) | |
|---|---|---|
| What it balances on | IP address and TCP/UDP port | URL path, hostname, headers, cookies, method |
| Protocols | Any TCP or UDP service | HTTP, HTTPS, HTTP/2, WebSocket |
| TLS handling | Passthrough to the origin | Terminated at the balancer, optional re-encrypt |
| Sticky sessions | Source-IP hash | Cookie-based, survives a changing client IP |
| Content-based routing | Not possible | /api to one pool, /static to another |
| Overhead | Lowest, packets are forwarded | Slightly higher, requests are parsed |
| Typical use | Game servers, databases, mail, SMTP | Websites, storefronts, REST and GraphQL APIs |
Balancing a web app or API? Layer 7. Balancing anything else? Layer 4. Running both? Configure both on one balancer. For application-layer DDoS protection and WAF, add Shield in front.
You register your origins and point your domain. We own the rest, algorithm, probes, certificates, capacity and the 3am failover.
Configurable HTTP, TCP or custom probes on every origin, with tunable intervals and thresholds so a slow origin is caught before a visitor notices it.
A 120+ PoP anycast footprint across four continents puts the balancing decision close to the visitor rather than close to your servers.
Free certificates issued and renewed automatically at the balancer, with optional re-encryption to the origin for end-to-end encryption.
Connection reuse and keep-alive pooling to your origins mean the balancer usually removes more latency than it adds.
Live per-origin traffic, connection counts, response times and health state, plus alerting when a pool degrades.
Real engineers on call. Failover is automatic, but when something needs a human, there is one.
Every plan is fully managed and includes both Layer 4 and Layer 7 balancing, SSL termination, health checks and DDoS protection.
Two origins and a real failover story
$20.00 Save 17%
Production HA for apps and APIs
$45.00 Save 17%
Multi-region, active-active, contractual SLA
Quoted to your traffic and regions
Prices shown in USD. Balance cloud VPS, managed VPS or colocated hardware, or origins hosted anywhere else.
Add a second origin, put a managed balancer in front, and a failed node becomes something you read about in a report instead of hearing about from customers.