Can you tell me more about NodeBalancers?

How do NodeBalancers work, and what does the service include?

3 Replies

A Nodebalancer listens on a public IP address for incoming connections, then uses configurable rules to select a backend node to which to send the connection. With a NodeBalancer, you can spread the incoming load to your application over any number of Linodes.

We have a guide which introduces NodeBalancers and gives an overview of how they work:

Introduction to NodeBalancers

When you have a website or an application, one of the most important things to consider is its availability. There's a term - High Availability - which is used to describe the setups of your server that eliminates single points of failure by setting up redundancy, monitoring and failover.

Our NodeBalancers are an integral part of creating a highly available website or application as it will distribute traffic between applications servers to make sure that none of them becomes overloaded.

Host a Website With High Availability

Once you have the NodeBalancer set up on your account and configured for High Availability, there are heath checks it performs, both passive and active. The passive monitors incoming requests for failure to connect, time outs, or certain 5xx errors and if it comes across any of those, it is considered unhealthy and taken out of rotation.

The active checks are proactive, and they run on intervals that are set up by the user. There are different settings for these active checks, and they are discussed in the guide below:

NodeBalancer - Health Checks

Here are some other helpful links on NodeBalancers and High Availability:

NodeBalancer reference Guide
Introduction to High Availability

Here is a bit more information on the stability, availability, and principles behind our NodeBalancers.

Principles of NodeBalancers:
The following section of our guide provides the best overview of how NodeBalancers work, and what they are based on:

NodeBalancers have a 10,000 concurrent connection limit. It's not a request/sec limit. There is no artificial request/sec limit built into NodeBalancers. A NodeBalancer config in TCP or HTTP mode can accept connections pretty much as fast as packets can be slung to/from the backends. In other words: it's a lot.

A NodeBalancer config in HTTPS mode can achieve 10,000 concurrent connections, too – it may just take some time to ramp up to that. While testing very small requests (connections don't live long) we've seen about 150 req/sec via HTTPS mode. Again, it's a good place to start, and we'll be working on improving the req/sec throughput of native HTTPS mode.

You can read more about this on our blog post

Stability and Availability:
A NodeBalancer runs in a clustered, high-availability, failover-ing, load balanced, etc, configuration. In the unlikely event that your NodeBalancer fails, the Linode it is running on fails, or even the host that is running that Linode fails, another member of the cluster will pick it up transparently within seconds. That said, if we are facing a datacenter outage, this would affect NodeBalancers and the Linodes they are pointing to. All of this would be covered under our 99.99% uptime guarantee.


Please enter an answer

You can mention users to notify them: @username

You can use Markdown to format your question. For more examples see the Markdown Cheatsheet.

> I’m a blockquote.

I’m a blockquote.

[I'm a link] (

I'm a link

**I am bold** I am bold

*I am italicized* I am italicized

Community Code of Conduct