Application performance SLAs. Every enterprise organization wants them. Few have them. Why is that? I’ve developed some insights from my role as Director of Sales Engineering at Aternity. Every week I perform dozens of demos of Aternity to IT organizations and lines of business in enterprises of the Global 2000. During these demos, I always ask the following question about application level SLAs:
‘Do you do performance-based service level management for your business critical applications?’
While you might think the answer to this question is an easy ‘yes,’ it’s not so simple. My question means more than just tracking application availability or up/down status. Every organization tracks availability. What I mean is setting performance SLAs for the application activities that support the enterprise’s critical business processes.
For a retail bank, this might be looking Customer Service Deskup a customer account in the teller app. For a healthcare provider, this might be accessing a patient record in the EHR. For a retailer, it might be scanning the bar code of the products at the POS in the check-out line. For application level SLAs, organizations need designate these business activities then monitor their performance as their workforce executes them in the client UI of the application.
Once I clarify my question, 95% of my enterprise prospects answer ‘no.’ They usually qualify this with statements like, ‘We’re thinking about it’ or ‘We’d like to.’
So how is it that it’s the year 2015 and so few of the organizations I talk to do application performance service level management? When digging in, I find that organizations wrestle with two fundamental challenges. With the right product, these are both easy to address.
Establishing Reasonable SLA Targets for Business Activities
Organizations struggle with establishing SLA targets for these business activities because they have no idea what the ‘normal’ performance for those activities is. How do you set a service level without know what normal is? For that small percentage of organizations who have said they are tracking these SLAs, they’ve usually tried to establish what’s normal by either doing a small amount of stop-watching timing in QA or Prod, running a synthetic script, or in many cases doing a search on the web to find articles on what service levels are acceptable.
The first two approaches usually fail because the amount of data collected is not statistically significant enough for Enterprises with tens of thousands of employees. The last approach usually fails because most SLA articles are focused on B2C web pages, and the typical 2-5 second ‘acceptable’ SLA values that these articles talk about are often completely unreasonable for a Search or Save activity that naturally could take 10, 30, 60 seconds or more based off of its functionality. Setting a 5 second SLA for these would put them in breach all the time and thus, they’re not reasonable. Only by truly knowing what is normal for these activities across the Enterprise, can a reasonable SLA be set.
Tracking Adherence to SLAs for from the Perspective of the Enterprise Workforce
Organizations which have set SLAs often monitor them via synthetic scripts. But monitoring via synthetic scripts defeats the purpose of an SLA. SLAs are meant to ensure the workforce sees good performance, not a robot. Without being able to monitor from the perspective of the workforce end user, it’s no wonder that enterprises don’t do application performance SLA management.
Two Key Steps to Application Level SLAs
Addressing these challenges requires two fundamental capabilities:
- The ability to track business critical activity usage across all users, seeing the performance (response times) every user sees in order to truly understand what’s normal. Only then can reasonable SLAs be set.
- The ability to then track the performance of every instance of these activities performed by the workforce, as they see it, and compare it to those service levels.
So how does Aternity solve this? Well, with our Business Activity Analytics, we monitor user experience in a very unique way. Aternity has the ability to monitor the performance of your business critical activities for every single user as they actually see them. Imagine you could stand over every user in your environment and stopwatch time every instance of them executing those activities. That’s effectively what Aternity is doing.
By doing so, Aternity can show what normal performance is for all your users across the Enterprise1, solving the first problem. Once this information is used to set reasonable SLAs2, Aternity will then monitor every instance for every user of these activities comparing them to the SLAs3, allowing you to now do Application Performance Service Level Management for your Enterprise4.
Application Performance SLAs, Whether or Not You Control the Apps
The great thing about monitoring end user experience from the perspective of your workforce’s device, is that you get insight into application level SLAs both for those applications you run from your own data center, and for those apps which may be run by a cloud provider, SaaS vendor, or outsourcer. As I mentioned in a previous APMdigest blog on End User Experience Monitoring, the approach that Aternity takes has advantages over other methods which require access to the app in question. As such, Aternity is used by a variety of customers for application performance SLAs for the entire range of apps, whether or not they are run in-house.
See how Aternity can help support your enterprise. Request a free evaluation today!