This is my first time posting. I'm hoping someone whose already been through this can help me out. I need to create a high-level web dashboard intended for executives as an at-a-glance health summary.
We organize our System Profiles as departments, so my first inclination is to make columns for each system profile that detail their health over 3 different time frames: 1 hour, 6 hours, and 24 hours, as well as an overall response time chart.
The main issue I'm having is determining what measure(s) to use (and how to configure them) for the traffic lights to accurately reflect the performance health of all the applications associated with the system profile. Ideally, I'd have a similar dashboard that would show each application individually as columns instead of the system profile, but with the web dashboarding tile limit, I've found it's difficult to get all of the applications on one dashboard.
I'm currently using the "Failed Transaction Percentage" measure, and I have it set to "Average" aggregation.
I don't think this is right though. The last hour tile constantly changes to red when there are seemingly no issues with the application, and the other timeframes stay green. I imagine that's because the average failed transaction percentage is skewed by the shorter period of time (1 hour) than the other timeframes, and also because the Web Page Requests Failed Transaction Percentage measure has no upper/lower thresholds defined:
I'm also unsure if I should be using the "Focus on the most actual value" or splitting based on the applications..
I think I have the right idea direction wise but I'm just not using the right measures or configuring the measures the right way.
Could anyone give me some advice on how I should be tackling this?
I think you should create for example business transactions splitted by application, and there you will have failure rate available to chart. Another things are incidents that can also be fired by your measures filtered for particular applications. If those applications are talking with eachother, keeping them in separate system profiles is not good choice 🙂
Hey @sebastian k.,
Thanks so much for the response. Could you let me know if this is what you're suggesting:
- Create a business transaction around an action that the application provides a lot of business value from (like a "Check Out" or "Submit" action.)
- Create a measure for that business transaction that measures the failure rate.
- Use that measure for an "Application Health" traffic light to indicate when the application is unhealthy.
If that's accurate I just have a few more questions to pick your brain about!
1. I'm having a hard time trying to think of a single business transaction that would reflect the health of the application accurately. Is it possible to create a measure for the failure rate of multiple business transactions instead?
2. Could a measure also be for business transactions of multiple applications, so we could have the overall health of the system profile being measure by the failure rate of all the business transactions we care about for that profile's various applications?
3. We have been categorizing our System Profiles kind of by department, but we have login processor applications that are used by all of our applications across System Profiles. Do you recommend handling that a different way?
Thanks so much for any more help you can lend!