“What URL’s or user journeys should I monitor for my website(s)?” This is a question we regularly get asked by customers when they are looking to sign-up for our external monitoring services.
The answer depends on a number of factors so we thought we’d share some of the rules we use to help customers decide what’s “mission critical” for their monitoring needs.
Rule #1 – “Ask your Marketing department or whoever runs your web analytics what makes money”.
The reason for this is that they will have mapped out the core “conversion paths” or “goal funnels” for your business. These conversion paths should be the business reason your website exists and if they are unavailable your website is losing money, hence it’s critical that you should monitor them!
You can see an example of a Goal Funnel as visualised by Google Analytics to the left of the page.
Rule #2 – “Do you have any major areas of functionality on the website that attract lots of traffic (even if they don’t make you any money per se)?”
A good example of this would be the RAC Route Planner – the route planner attracts a lot of traffic to the RAC website but it isn’t a “money maker” as such when compared to the e-commerce user journey to buy RAC membership (which would be covered by Rule #1).
However the route planner is a very high profile application and is used by RAC members and non-members on a daily basis. If the route planner were to go down it would hurt the RAC’s brand and reputation, which are worth investing to protect.
Rule #3 – “Are different areas of your website hosted on different web farms or other infrastructure?”
It’s very common to split your application into different domains, sub-domains e.g. payments.mycompany.com or to use your load balancer to split sub-directories on to different web farms e.g. /payments/ might send traffic to a more secure PCI compliant server farm compared to /products/.
If these different sub-domains or sub-directories can fail independently of the main www.mycompany.com domain then they need to be monitored separately otherwise you can be lulled into a false sense of security that just because www.mycompany.com is up so is the rest of your website.
Rule #4 – “Do you have any special ‘status pages’ that can give insight into the health of your website?”
A great way to get more “back for your buck” from your external monitoring is to ask your developers to build a “status page” that behind the scenes checks the health of core parts of your application e.g. can it still connect to the database, the payment gateway, the search server etc.
Each separate part of the application can be reported in a table or text on the status page. If the status page reports a service as “down” the page monitor can detect the string “down” in the HTML and trigger an alert. The status page can also be used internally on a large screen display to help alert your operations when problems occur or when they received an alert from the external monitoring so they can “at a glance” know what’s going wrong.
Rule #5 – “Are there other services your website depends upon other than just plain HTTP web traffic from your own web servers?”
A modern website is more than just HTML served over HTTP/S from a web server.
Websites rely on SMTP to deliver emails for customer registration, purchase confirmation or re-acquisition targeting. It might rely on XML or RESTful webservice provided by 3rd parties for services like postcode lookup, address verification, quotation engines, price comparison or customer reviews. It might even still be using “legacy” services such as EDI document exchange via FTP.
You need to consider the wider ecosystem in which your website operates and how failure within that ecosystem might trigger a collapse in your own systems. If external monitoring can give you early warning of failure or help you rapidly troubleshoot the problem then that’s well worth considering.
If you do rely on 3rd parties it’s well worth thinking about the SLA you have with that provider and how you’re verifying that they are living up to their end of the bargain. Are you relying on them to monitor their own SLA’s, a breach of which will probably trigger penalties and cost them money?
In that scenario a “trust but verify” approach using independent external monitoring might end up actually saving you money if you find that their actual performance is significantly lower than that which they report to you!
We hope these 5 rules help you in your own monitoring strategy and if course if you want to talk to any of our Professional Services or Account Management team to learn more just give us a call!