Downtime is one of the most worrying circumstances a website owner can face - loss of access to a service can result in customer dissatisfaction and mistrust and can result in unknown time costs and frustrations for a site owner. Rand Group estimate downtime costs for a single hour can hit [hundreds of thousands in revenue] (https://www.randgroup.com/insights/cost-of-business-downtime/), a number that rises year on year.
Many business owners come to us for help with existing sites that are going through that dark period. When faced with an unknown project where delivering full functionality is a priority, we go through a process to ensure we can solve issues quickly, effectively, and without further impacting the end-users.
Prior to a deep investigation, we seek to mitigate any existing downtime by treating symptoms of the problems. Our development team has the existing knowledge with common toolkits and content management systems such as WordPress, Umbraco, phpBB and more, to allows us to quickly find mistakes or security flaws that previous developers might have left behind.
We set up monitoring and diagnostics that let us examine the health of a solution. This provides us with valuable information regarding CPU usage, memory usage, network transfer and more that alerts us to any issues without delay, and often allows us to predict downtime before it happens. We provide this information to site owners through a web interface, so you aren't left in the dark until it's too late.
Once diagnostics are set up, we replicate the deployed servers for our own testing. With these newly set up machines, we can make configuration changes and compare their health to that of the live system. We experiment with ways to provide a full solution without ever causing additional negative impact to the live service.
When we've found potential fixes, we continue to monitor systems overtime to ensure there's a full solution, not just a temporary patch. We want to ensure our fixes hold up over time whilst also being able to deploy changes as soon as possible. To do this we use tooling to stress test servers, sending hundreds of page requests to our test servers at once to simulate heavy load in a safe environment.
This also serves as an additional diagnostic process, revealing any further issues before the live site experiences them.
Once we're content that our changes are stable, we prepare to deploy the changes to the live system. We work with the client to identify a safe period of time for delivering the new site, making sure any content management or promotions won't be disrupted. We aim to deliver as transparently as possible to ensure end-users aren't adversely affected.
Following deployment, we continue to monitor the live service to guarantee the site stays stable and without issue. Our comprehensive diagnostics give clients peace of mind that issues won't go unmissed.