The way IT departments monitor business-critical applications and services is changing. It used to be that network, server, and infrastructure monitoring by itself was enough to give you the service assurance you needed. Now, with more businesses relying on bespoke, custom-coded applications, the old way of working doesn’t give you the whole picture.
The need for a more application-centric view of IT infrastructure performance has given rise to application performance monitoring (APM) solutions. APM augments traditional IT service monitoring, providing businesses with a in-depth view of how all their applications are performing for customers and employees. But is APM always a requirement? Is it a solution for all your IT monitoring needs?
That discussion is at the heart of our recent TalkingLogic podcast, where Russ Elsner, from the Office of the CTO at ScienceLogic, discusses the importance and implications of APM.
“APM is all about code-level transactions, tracing, application mapping, and seeing how your code is performing in your production environment.”
Russ Elsner, ScienceLogic
What’s the Difference Between APM & Traditional Service Monitoring?
It’s easiest to think about the contract between traditional IT service monitoring and APM as being about breadth and depth.
Traditional IT Service Monitoring Gives You Breadth
Think about standard IT service monitoring—tracking and managing network availability, server uptime, database capacity, response times, and more. This type of monitoring lets you know how your services are performing as a whole. It’s a broad approach—you know that if your network is slowing down due to bottlenecks, it’s going to impact every service and every user on that network.
Application Performance Monitoring Gives You Depth
In contrast, APM lets you look deeply into how a specific application is working. APM is typically deployed into the operating system on an application server, where it tracks, monitors, and reports on how the custom code there is performing.
For example, say you’re a travel agent who lets people make bookings using a custom-written ticketing application. APM would sit on the server and report on how long it’s taking customers to get through the system—are there bottlenecks or slowdowns in the database? How are the integrations with airlines or hotels functioning? What are response rates like for user queries? APM can then tell you where code is slowing down or failing, so you can get your development team to deploy a fix and preserve your revenue.
Taken together, traditional monitoring and APM give you a complete picture of performance for your most important apps. The traditional monitoring means the foundations of your IT are performing well, whereas APM tells you that the house you built on that foundation doesn’t have a leaky roof.
“If you’re hiring your own application developers to create custom solutions for your business, APM is something you can’t do without.”
Russ Elsner, ScienceLogic
Does Every Business Need APM?
No, not every business benefits from APM. If you’re mainly using off-the-shelf, locally installed, SaaS or cloud-based apps written by others, you typically won’t benefit from it. Consider commercially available applications from SAP, Microsoft, Salesforce, Oracle, and the like. If you were to detect a code-level issue that was impacting performance, what could you do about it? Nothing! And at a cost of $150–$200 per APM agent, it doesn’t make financial sense to deploy APM across your broader IT services.
However, if your business has developed its own custom apps, then APM could be the solution you need. Consider using APM if:
- Your business develops custom IT application solutions from scratch
- You have lots of systems that interact with other custom IT solutions
- Your IT custom application is a major revenue generator or critical part of your business process
- Your application is out of regular vendor support and you rely on internal IT teams to support it
These are all good reasons to get APM in place. It can give you the reassurance the application itself is running at peak efficiency.
If you’d like to learn and understand more, please listen to our latest TalkingLogic podcast and hear Russ break down and explain what you need to know about APM.