cancel
Showing results for 
Search instead for 
Did you mean: 

What Does APM Mean to You?

Level 10

Many of us have heard the term APM, and just as many are confused as to what it truly means. Gartner put together their own definition – you’ll find it here. When writing this post I was thinking about 2 of the 5 “functional dimensions” Gartner outlines as making a proper APM solution. These are:

  • Runtime application architecture discovery modeling and display: As I like to think about it, this means discovering and providing useful metrics on the full range of paths that an application can take between software and hardware in your environment as part of proper execution.
  • User defined transaction profiling: Similar to the above, but more focused on how real users are using the application, and thus the paths their actual application requests take through the same hardware and software topology. Think about it like choosing those paths which are most critical to your actual users. This allows the solution to provide real-time metrics surrounding user experience and satisfaction.


We all have a number of web based apps running in our IT world, many of whose topology we may not fully understand. Capabilities such as the above can be helpful in identifying, troubleshooting and isolating the cause of user issues with those apps. After all, the issue could be anywhere in the various layers of our environment, and no one wants to start a guess and check game of servers, databases, etc. you stand to lose while wasting valuable time. At the same time, we’re all curious how exactly users are interacting with our environment, but don’t possess the sixth sense to tell us just which DB has the most users hitting it at any given time. With those two capabilities combined, you can begin to imagine being given visibility into which issues are impacting the most users at any given time (thus where the most helpdesk tickets will come from) as well as the components along their application request path that could be the root cause – all in real time.

In thinking about all the above, I’m curious about the following:

  1. Do you or your company currently use any software today that helps provide this sort of information?
  2. If you do, what sort of problems has it helped you to solve? With what type of applications? What do you feel is missing?
  3. Would knowing your various applications’ topology be interesting to you or your company? How about the real user transaction paths within those same applications?
  4. What sort of problems do you think could be solved if you knew all of that?
  5. In general, do you wish you knew more about how your real user’s actions affect the IT environment you work hard to monitor and maintain?

Don’t be shy and comment! I know I’m not the only one that struggles with this and, after all, misery loves company.

10 Comments
clubjuggle
Level 13

I think we all wish our software could proactively read and model users' actions, but I'm not sure how that would be accomplished. As it is, we as APM platform managers need significant input from the application support people and/or power-users, and possibly also to mine the ticket history for signs of recurring problems, in order to determine the means to detect those problems proactively going forward. Those tickets can give insight into actual user behavior.

jkump
Level 15

We don't use any solutions but it would be great to gain an insight and understanding into the world of the user and as a computer scientist I would relish the capabilities that could come from such modeling.  However, time will tell.

Jfrazier
Level 18

Modeling user action is a challenge.  Some people do strange things at times for no apparent reason.

On that same note..User defined transaction profiling can be somewhat accomplished.  There are challenges when it comes to certain types of transactions.  In some environments, performing those transactions costs money.  In some cases, you can't build a "dummy" transaction  to test or validate the process so you are only seeing some aspects or a partial view of user environment.  In a previous life we were able to set up a "fake" order of parts to simulate a customer online experience..while it was like 2 or three items, at the surface it validated the environment was up and running.  Currently the challenges are with loan products and the many things that tie into a successful application all the way through funding.  We are able to test some aspects via WPM, but others to outside vendors cost per transaction or we would have to select live apps.  That starts to get sticky as we don't want to see customer info and all the regulations surrounding that.

cahunt
Level 17

We have and have had SAM but use it in a limited manner. We are looking forward to better collaboration with other departments and the applications teams to expand this.

Knowing the topology is always key, as it is our distributed server setup and Storage Array's always create a bit of an issue in tracking when the 'Service' goes offline for a specific application or process.

 
App teams sometimes have to check each box to find the service that has hung up or stopped - and this is not always visible off the top. Having a tool that would leverage the network connection and information with throughput of data along with the reference to server resources and service connections would be the key in knowing what disconnected or what service stopped to cause the application issue.

jkump
Level 15

Sign me up when you get it!     Fight those same battles here.

byrona
Level 21

To me its the Holy Grail; many have searched for it and nobody has found it.  I have seen products and services that claim to do this but in the end they all seem to do it in a limited capacity.  It seems to me at this point it's more cost effective to monitor the underlying components of the application and the output of the application and this will let you know if it goes down or is having problems and where to start looking.  From there it all goes back to traditional troubleshooting.  Basically this is what you can do with SAM now; then again, maybe I am just a traditionalist. 

jpashton
Level 10

We have several in house programs and do use SAM, but we mainly use it for monitoring of the corporate server infrastructure, as the maintenance and monitoring of the in house programs is primarily the job of the programmers here.  We just monitor server, network, and storage status of the various in house builds.  We mainly use the SAM for monitoring the SQL and Exchange servers as well as the other various application servers under the IT department's wing.  I'd like to be able to go more in depth with some of our monitoring, but personally don't have the time to go as in depth as I'd like and I'm as close to an "expert" on all of our SolarWinds monitoring tools as the department has. 

cahunt
Level 17

Indeed. I am looking forward to helping folks figure some issues out once a full view is in place. Already looking forward to new 'Saved Bacon Events'.

ecklerwr1
Level 19

I thought APM was the old name for SAM... I use SAM.  It works pretty darn well and even if there isn't already a template made you can either make one yourself or ask

aLTeReGo!

Seriously that was just props for alterego... he's done a great job with SAM.  Bronx too btw.

User profiling from all of your remote sites is kinda what we'd all like to have but it takes a lot of work to do right.

omri
Level 10

Hey jkump - what sort of capabilities do you think would come from such modeling and how do you think it would help your day to day?