There’s been a spate of high-profile website outages over the last few months. Social media platforms—including WhatsApp, Facebook, Twitter and Reddit—have all been guilty, not to mention travel sites Ryanair and BA.

In addition, Amazon Web Services took down half the internet when an update went awry, which might have been down to a typo. However, typically it’s the failure of businesses to properly monitor and test their websites that causes unexpected and unacceptable downtime. Many aren’t aware of how much traffic their site can handle, how to prepare for busy periods, or how to respond to an unexpected surge in users. Until it’s too late, that is.

>See also: The four ways to deal with web application security: which is right for you?

To put a stop to outages, which drive away customers, damage reputation and lower market confidence in a provider, organisations have to invest in proper monitoring and testing services. And like a good suit, monitoring should be tailored depending on need.

Different sites offer different services; they need to be treated as individuals in order to identify specific weaknesses and create an action plan. Here are three points to keep in mind for testing and monitoring your website.

Audience

Every website has a specific purpose: a website that sells gig tickets and a niche blog simply appeal to different scales of users. The gig ticket website will need to be able to cope with large bursts of website traffic, whereas the personal blog will have much smaller peaks.

Stress testing lets you see exactly how much your site or app can handle before falling over, and therefore whether you need to increase your server capacity. Comparing the absolute bottleneck with typical traffic lets you stay on top of potential failures.

Key dates

Some websites can predict when their busy periods will be. Amazon knows it will be inundated over Black Friday, whereas PC World will see peak traffic on Cyber Monday. Consequently, there is a specific calendar that they adhere to when it comes to website traffic.

>See also: Why penetration testing is a vital part of any effective security strategy

Concurrency testing allows someone like a major retailer to see how many users their site can handle at any one time, so when their site is needed most, it can continue to function at its best.

Curveballs

Unexpected surges in traffic are by their nature difficult to predict. As recently seen with the FFC going down when John Oliver asked viewers to visit their website, influencers and promoted material can seriously affect sites and applications. If a company feels their website could be particularly vulnerable to this sort of ‘flash traffic’, then scaling and disaster recovery testing is a smart choice.

In the same way that off the peg will never be as good as tailored, monitoring does its best work when it’s specific to the needs of a website. It is important to have the right tools in place to monitor website traffic and test out your products, so any issues that arise can be rectified quickly and responsibly.

Sourced by Erik Torlen, CTO of Apica