[go: up one dir, main page]

ACHIEVING HIGH AVAILABILITY

Five nines is no longer a myth.

haproxy high availability

Fault-tolerant availability has been built into our design since the very beginning. Get closer to 99.999% uptime with the HAProxy suite of products.

Key Principles

graphic image of a laptop with 24/7 written on the screen

High Availability is the principle of keeping a website’s content accessible throughout the year. An uptime as close to 100% as possible is today’s chief aim for any company striving to perfect their web operations.

high availability layers

To achieve high availability at each layer of your infrastructure, from the network edge down to the application level, our products tackle the complexity for you.

haproxy logo in a green, blue and red circle

With a suite of in-built features contributing to this goal as well as a centralized hub for managing clusters of load balancers, HAProxy Technologies load balancing solutions bring absolute availability closer than ever.

Learn how to achieve high-availability with HAProxy

HIGH AVAILABILTY PLAYBOOK

The Fight to Remain Available

Maintaining this high uptime means fending off several key threats. Faced with the challenge of staying online, companies should be on the lookout for the following common weaknesses.

broken chain icon

Network Design Flaws

SOLUTIONS:

  • Failover

  • Advanced Health Checks

  • Retry Requests

  • Circuit Braking

  • Support Tools

A well-designed system is the first line of defense against downtime. Minimizing failure-points by running a website across several servers is key, with a seamless backup plan for outages like HAProxy Enterprise’s automatic failover settings when one goes down.

But keeping a network healthy is always preferable to a server taking a sick day. In order to ward off these interruptions before they strike, benefit from the advanced health checks offered by the HAProxy suite of products to continually monitor your servers for downtime and errors. And when components do fall under-the-weather, support for retry requests on other servers as well as circuit breaking means that traffic can be routed around unhealthy server groups during longer outages.

The same principles apply for the load balancers themselves. By creating a cluster of load balancing instances you enable redundancy in case one fails, and with support tools for an Active/Active or Active/Passive setup, your backup load balancing instances will always be ready. And with the HAProxy Fusion Control Plane you can fluidly manage the setup and configuration of these multiple load balancers all in one place.

graphic image of white bricks on a blue background

Security Attacks

SOLUTIONS:

  • Web Application Firewall (WAF)

  • PacketShield

The internet can be a dangerous place. Given load balancers are often the first line of defense against unwanted visitors, taking advantage of a sophisticated security system like the HAProxy Enterprise Web Application Firewall to scan requests for malicious payloads as they arrive will allow you to stop threats before they can reach your servers and disrupt your uptime.

Defending against denial-of-service threats can necessitate a different response. A network layer solution like HAProxy ALOHA’s PacketShield is another approach, and ideal for stateful packet filtering of illegitimate traffic.

deployment risks icon

Deployment Risks

SOLUTIONS:

  • Slow Start / Slow Stop

  • Hitless Reloads

Continuous delivery can mean that application deployments happen frequently, even several times per hour, meaning that updates must be carefully planned so visitors can continue to be routed through as services reboot.

Taking advantage of HAProxy Enterprise’s slow start and stop will steadily increase and drain traffic to servers as they undergo this process. Hitless reloads also allows seamless updates on the Load Balancer itself – meaning that it can continue to route traffic as configurations are updated.

fine tuning icon

Fine-Tuning for Strength

SOLUTIONS:

  • Monitoring

  • Traffic Control

  • Timeouts

  • Connection Pooling

Round out your strategy for high availability with our range of other features for monitoring and traffic control. With the most detailed metrics of any load balancing solution and robust support for third-party monitoring systems, HAProxy Enterprise helps you pinpoint your errors to fix them as they occur.

Additional features like timeouts and connection pooling also mean that your servers are always running at maximum efficiency, alongside other options like traffic overload protection to keep each load at a manageable level with queues.

Contact the authoritative experts on HAProxy who will assist you in finding the solution that best fits your needs for deployment, scale, and security.

Contact Our Experts