A load balancer takes every request for your site and shares it between your cloud servers, checking each one as it goes. A server that stops answering is taken out of rotation until it recovers, so a single failure is something your visitors never see.
Requests are shared between healthy servers. Add another server and it joins the rotation; take one out and nothing else changes.
The size sets how many services and servers a balancer can handle and how many connections it can hold open. You can move between sizes at any time.
Two or three servers behind one site, the usual first step away from a single machine.
Several sites or environments behind one balancer, with room to grow the pool.
Large pools and busy applications, with the most open connections and the most certificates.
Every balancer comes with a public IPv4 and IPv6 address, and a private IPv4 when it joins one of your private networks. Prices are per month; annual billing takes 10% off and is charged once for the year. Prices exclude taxes where they apply.
The pool is made of our unmanaged cloud servers: root access, NVMe disks and a choice of three lines in Europe, the United States and Singapore. Start with two of the same size and add more as you grow.
The lowest price on WebSols Cloud, on processor generations that have been in service long enough to have nothing left to prove. The disks, the network and the traffic are the same as every other line; what you share is the processor, so speed moves a little with the neighbours. Ideal for test boxes, staging copies, prototypes and small sites.
Newer AMD generations, still shared, with the best balance of price and speed. The usual choice for websites, small databases, staging and development.
Dedicated vCPUs, one per physical thread, so nothing else competes for your processor time. For busy databases, game servers, CI, SaaS and anything under constant load.
Servers and balancers appear on one monthly invoice, in rupees. A balancer and its servers must be in the same region, though they may sit in different data centres within it.
See cloud servers and pricesYou do not need to change how your application is built. The balancer sits in front of what you already run.
The balancer gets its own public IPv4 and IPv6 address. Your DNS points at it instead of at a single server, and nothing else in your domain changes.
Servers join by public IP, or by private IP when they share a private network with the balancer, so the servers themselves need not face the internet at all.
Requests go round robin or to the server with fewest connections, and health checks decide who is in the rotation. Failing servers come back automatically once they pass again.
Set up with you when the balancer is built, and changed whenever your application changes.
Share requests evenly, or send each one to the server holding the fewest connections. Pick whichever suits the application.
Balance any TCP-based traffic, or work at the HTTP level with HTTP/2 where the visitor's browser supports it.
A check every 3 to 60 seconds, with a timeout of 1 to 60 seconds and between one and five failures before a server is pulled out. HTTP checks can test a path, expected status codes and a string in the response.
Terminate TLS on the balancer and take the decryption work off your servers, or pass encrypted traffic through untouched with a TCP service.
Upload the certificate you already hold, including a free one, and we install it. Automatically renewed certificates are available depending on where your domain's DNS is hosted, so ask us.
Keep a visitor on the same server with a cookie, for anything between 30 seconds and 24 hours, when your application needs it.
HTTP and HTTPS services add X-Forwarded-For, X-Forwarded-Port and X-Forwarded-Proto, so your logs and your application still see who is asking. The PROXY protocol is available for TCP services.
The balancer itself is redundant. If the hardware under it fails, another takes over automatically with only a short interruption.
Move up a size when the pool grows, or back down again. Services, servers and connection limits change with it.

Once traffic is shared, growth is a matter of adding another machine to the pool, which is cheaper than doubling a single server and safer than relying on one. It also makes maintenance boring: take a server out, patch it, put it back, with nobody watching a maintenance page.
What each size holds, at a glance.
| WS-LB1 | WS-LB2 | WS-LB3 | |
|---|---|---|---|
| Services | 5 | 15 | 30 |
| Servers in the pool | 25 | 75 | 150 |
| TLS certificates | 10 | 25 | 50 |
| Open connections | 10,000 | 20,000 | 40,000 |
| Traffic included, Europe | 20 TB | 20 TB | 20 TB |
| Traffic included, US and Singapore | 1 TB | 2 TB | 3 TB |
| Health checks | Active and passive | Active and passive | Active and passive |
| Sticky sessions | Yes | Yes | Yes |
| Public IPv4 and IPv6 | Yes | Yes | Yes |
| Automatic failover | Yes | Yes | Yes |
A service is one listening port and the rule behind it, so a site on 80 and 443 uses two. Servers in the pool may sit in different data centres, as long as they are in the same region as the balancer.
Load balancers sit alongside our unmanaged cloud servers, so the split is the same one: the platform is ours, what runs on your servers is yours.
Individual requests are not logged, for privacy and performance, so the balancer cannot produce a per-request access log. Your own servers still log everything they answer.
The unmanaged NVMe servers that sit behind the balancer, with root access and monthly pricing.
Read more →The same servers, run for you, with a control panel, security, backups and 24/7 support.
Read more →Autoscaling shared hosting, when a pool of servers is more than you need.
Read more →What people ask before putting one in front of their servers.
When one server is no longer enough, or when you cannot afford for one server's failure to take the site down. Below that, a single server with backups is simpler and cheaper.
When one server is no longer enough, or when you cannot afford for one server's failure to take the site down. Below that, a single server with backups is simpler and cheaper.
Tell us how many servers you run and what they serve, and we will size the balancer, set up the health checks and install your certificate.