Do you trust your critical infrastructure?
Written by David Mytton
The infrastructure behind any company has multiple components; it’s the tools and systems that allow the business to run. You might think that every part is equally important, but that’s not really the case. You can divide your infrastructure into at least 2 categories – critical and non-critical.
The former contains things which if they fail, would cause “downtime” in the sense of loss of productivity, revenue or impact on reputation. This could be the internet connection to your office failing or your web hosting provider having a data centre outage which takes your website offline.
The latter are things which are important, but don’t stop business as usual. This could be a fridge breaking down in the office kitchen or the tool you use to manage your Twitter account going down.
For every component in your infrastructure you need to have a plan to determine what happens if that component fails:
- How will you find out if something is broken? Are you using a monitoring tool like Server Density, or will you find out only when your customers are complaining? Even then, do you monitor support out of hours to check for reports of critical issues?
- How critical is this component i.e. when it breaks, what priority should it get in relation to your normal workload?
- Who is affected inside your company when this item breaks? Is your entire team unable to access the internet or does it just mean there’s no convenient place to store your salad for lunch?
- How do you get support for the failed component? Do you just have to wait for an e-mail response or is there phone support? What are their opening hours? Do you have a response time SLA?
And perhaps the most important thing for your critical components: have you tested the support from your vendor? If your payment processing breaks at 11pm on a Saturday evening, can you call someone to investigate or will there not be a response until Monday? Does the SaaS provider that handles your e-mail have a method to reach them when your e-mail is broken? You don’t want to be discovering these limitations when something isn’t working and you’re losing signups/revenue. Don’t just assume, test your vendors!