Who hasn’t clicked on a web page and received an error message? The consequences range from mildly annoying (can’t get the information you need) to hugely consequential (a system crashes during a big financial transaction). Maybe you click away, slightly irritated. Maybe you take a screenshot of the issue and email the company that owns the website.
Whatever the case, in today’s fiercely competitive business world you do not want to be the company in this situation. If your end users are the first to detect a problem, this has ramifications not only for that transaction and that relationship but potentially for your reputation and future sales.
The message is simple. Don’t risk poor monitoring of your technology infrastructure, platforms or applications. You must be proactive so that you detect any issues well in advance of them becoming a serious problem. Here are five indicators that you may need to lift your game.
1. You are experiencing regular downtime
Technology downtime should be the exception. In the hectic demands of day-to-day, talented teams can work around difficulties and even start to view them as par for the course: ‘The emails system is down again, send me the information on “slack/hipchat/skype” instead.’
This is a simple example that can cost big amounts in time and client relationships management. But the bigger your business, the more crucial your applications, the bigger the noise and the potential risk if things are regularly out of whack. One e-commerce platform that we look after generates around $1 million of sales per hour – so even a short-term glitch has a very big price tag.
2. You find out too late about failing or under-performing applications
This is linked to the scenario above. It may be end users that report a problem. Or, worse, problems can go undetected for some time, slowly causing a chain reaction of other issues that are not dealt with.
For example, a problem with outgoing emails this can be hard to detect. The email program itself won’t alert you. The clients or business partners not receiving your email won’t call to tell you they haven’t heard from you – unless it’s got to urgent stage, at which point your issue is out of hand.
Proactive monitoring means you are the first to know, not the last. This means that you must put your monitoring tools on different networks. Ones that will still be running if your infrastructure dies or other downtime or performance issues occur. Your monitoring must always be one step ahead.
3. You didn’t include monitoring in original set-up plans or budget
There is a misconception around IT that today’s backup systems are foolproof and that the people setting up your platforms are automatically thinking about monitoring.
Neither of these things is true.
It is up to you to ensure that proactive, comprehensive monitoring is scoped into every project. Don’t assume it will be, don’t wait to include it later and don’t skimp on it in terms of budget. Once the project is handed over to the operations team it’s too late to try and wedge monitoring in down the track – both in terms of system set up practicalities and the issue of cost.
If you don’t know whether proactive monitoring was part of the original scope in any aspect of your solution, it probably means it wasn’t – an issue you need to address promptly.
4. You have some monitoring in place but it is ad hoc and not thorough
‘Some’ monitoring is a more common error than none at all. Unfortunately it leaves you in just as precarious a position. IT platforms, services and systems operate at a number of different, yet interrelated, levels. You need to ensure each layer is being proactively monitored so that any problems can be either automatically or swiftly identified and resolved. Consider, how – and when – will you know if you have an issue with any of the following:
Infrastructure layer
Platform layer
Application layer
Business process layer
5. You have relied too heavily on developers and not been involved enough
It is tempting to project a kind of god-like knowledge and awareness onto development teams. Especially now as DevOps environments become more common, you may assume that a detail such as proactive monitoring will have been picked up or just naturally incorporated into the start of the project.
As good as DevOps is, it has its limitations (more on this in an upcoming post). The main ‘blind spot’ is that most projects concentrate on functionality and user experience and forget to seriously consider what happens if platforms or systems fail. The focus of testing tends be on functionality, with developers often forgetting about the need to test supporting layers.
This isn’t to place blame. It’s simply a reminder that you cannot assume proactive monitoring has been at the forefront of others’ priorities during development and implementation phases. And it doesn’t then present as an issue until something goes wrong.
If proactive monitoring is something you have been a little bit ‘head in the sand’ about, now is the time to come up for air. We can assist with any queries you have, from an assessment of your current set-up to recommendations about what needs improving.
Comments