How to Ensure High-Performing Microsoft .NET Applications

Editor's Notes

1. My name’s Natalie Tomko and I’ll be your moderator today...
2. ...here’s a few logistics before we begin...
3. Our speaker today is John Worthington, Director of Product Marketing for eG Innovations....so at this point I’ll hand the ball off to him...welcome John...
4. Thanks Natalie and welcome everyone....here’s our agenda. We’ll do a quick introduction to eG Innovations and an overview of the .NET framework to set the stage, and then we’ll get into a discussion of the why, what and how of full stack .NET monitoring...we’ll also give you a quick demonstration and take a few questions along the way.... ...and then I’ll summarize how converged application and infrastructure performance monitoring can help improve your ability to monitor .NET environments ... Then we’ll answer any final questions and wrap it up for today.
5. eG Innovations has been providing performance monitoring solutions for over a decade...our patented data flow and dependency-based analytics are at the heart of the product, and automate the diagnosis of performance issues at any layer of any tier of a digital business service– from code to bare metal and across any deployment model, public private or hybrid.
6. ...and this is why people come to us.... ........... So, let’s dive into how .NET fits into this picture.
7. Microsoft .NET is one of the most popular web technologies....millions of people use it to develop Windows client applications, web services, distributed components, client-server applications and database applications...
8. So, let’s quickly break down the Microsoft .NET architecture.... ...web sites, services and applications use .NET languages ...using ASP.NET framework to build the application and services. The common language runtime environment manages the execution of .NET code, including things like memory management and debugging... ...and the IIS is a web server that’s used to host the ASP.NET applications. <CLICK> the bottom line is that an issue in any one of these layers can effect the performance of a .NET application or service
9. ...and this is pretty important, since slow is the new down and end user experience has become the new basis of competitive advantage... https://devops.com/real-cost-downtime/ https://www.novell.com/communities/coolsolutions/much-slow-business-applications-costing-enterprise/ https://medium.com/@georgea88163107/8-principles-of-user-optimization-thatll-increase-your-search-rankings-139fa1f0df46#.iajqssjv3
10. Monitoring the user experience is the guidepost for performance monitoring for this reason...IT must remain focused on the digital user experience... ...and diagnose what is causing a slowdown, hopefully before it’s noticeable to the user... ...remember, the user will not know or care whether it’s due to the front-end, network, download or server....that’s NOT their problem...it’s YOURS... So user experience is key.
11. An important part of the digital user experience are those digital business transactions.... This is where transaction tracing comes into play.... ....we need to be able to quickly drill down and get visibility into the application code and/or a database query, identify the cause and get it fixed ...so monitoring business transactions is important as well
12. ...of course performance issues can be in other parts of the .NET stack.... ....they could also arise in the .NET framework or in the web server... ...so, of course we need to monitor the .NET application too
13. ...and finally, the .NET framework and web servers are supported by infrastructure...and issues here can also create performance problems... So we can’t ignore this layer either...
14. Full stack .NET monitoring needs to correlate end to end at every layer of the .NET stack and include... User experience, business transactions, .NET application and infrastructure performance layers....
15. Let’s take a look at each of these layers in a little more detail...
16. OK, at the top we have user experience....we can monitor this based on real users or we can perform synthetic user monitoring... A best practice is to use a combination of both of these approaches... ...let’s take a quick look at what real user monitoring looks like.... DEMO 1
17. So, there’s a lot that real user monitoring can do for you.. ...the visibility into the user experience can help you detect problems early and improve mean-time-to-resolve performance issues...the visualization and dashboards can help improve communication with business users and the alignment of technology performance KPIs with business critical-to-quality indicators.... ...the net result is that monitoring the end user experience can be a trigger for improving the quality of your digital business services....
18. ...but while we can see that RUM will show us that server-side processing is slow, <CLICK> it won’t tell us why....and effective application performance management is all about getting to actionable information, right?
19. ...the next layer we said was the business transaction layer... ...this picks up where real user monitoring stops .... Transaction tracing provides visibility across the server-side tiers of the .NET application. We trace transactions using a tag-and-follow approach all the way to the database and back, and it allows us to identify slow or stalled transactions ...or transactions that are having errors... ..we can also get the response time along each tier of a distributed .NET transaction so we know where the slowness is originating from....
20. Here’s how we do transaction tracing.... ...we use byte code injection....<CLICK> the injected code reports the execution time for each step of a transaction across the application architecture......<CLICK> and adds a globally unique identifier....this tag-and-trace approach allows us to follow the transaction across the entire path... We can trace through multi-tiered IIS servers and CLRs, support horizontal and vertical scaling...and we can also trace across hybrid .NET and Java architectures
21. I should also mention that this is easy to deploy and doesn’t require and special access or configuration....or any changes to your application code in order to trace transactions
22. By monitoring the .NET framework as a layer within an IIS web server component, including the .NET applications, worker processes and the .NET CLR... ....you can determine whether, cache, sessions, requests, wait times, garbage collection, heap memory....and so on...could be the source of a performance problem....
23. ... Monitoring the IIS web server includes... ...requests, TCP connection time, cache hits or misses, CPU, bad requests....any of these could effect the performance of a .NET application or service ...let’s take a look at this... DEMO 2
24. So if poor performance is caused by defects in the application code, a poorly designed query or slowness in remote calls... ...transaction tracing will identify the cause... Let’s pause and take a few questions.... How do I know if there is a problem in the ASP.NET framework or in the underlying IIS server infrastructure? The layer model we saw in the demonstration helps you easily distinguish between performance issues at different layers of the .NET stack.... Remember the pre-built monitoring model for IIS --- and the embedded analytics provided by eG Enterprise --- provides deep visibility into CLR exceptions, garbage collections, heap memory usage, worker process and application pool issues, IIS web cache, HTTP errors, and more. We have Java servers communicating to some .NET custom applications. Can eG Enterprise understand the dependencies between our Java and .NET architectures and trace transactions through both? Yes, we support tracing distributed business transactions through cross-platform .NET and Java applications. You instrument eG Enterprise in JVMs the same way you do it in .NET CLRs, and you can trace the entire transaction .... hop by hop across the Java and .NET server-side tiers... What machine learning and AIOps technologies does eG Enterprise support? We’ve used machine learning to automatically baseline any environment and pinpoint performance deviations for more than a decade now. The embedded analytics include a patented dependency-based performance correlation technology to discover dependencies between applications and infrastructure, automatically isolating the root cause of performance problems. What’s also nice is that the analytics are reflected in the built-in reports for capacity planning, forecasting and infrastructure right-sizing and in the dashboards, which provide what I like to call situational awareness... Every stakeholder has a real time, personalized view of what’s happening from a performance standpoint. Ok, let’s continue the webinar and take more questions before we wrap it up today...
25. ...but infrastructure issues also affect application performance, so in these cases transaction tracing will only point the finger there.... ...your IIS server and everything it supports could depend on physical and logical components....operating systems, hypervisors and other systems like active directory or storage.... ....in order to manage the performance of .NET based services, you need total performance visibility and the ability to automatically correlate all aspects of application performance...
26. ...so in order to be confident that you’ve identified the root-cause of a performance issue, you will need to monitor the infrastructure as well as the application.... ...if a backend is slow due to a lack of memory, then all queries are going to be slow.... <CLICK>transaction tracing may tell you that a query is causing slowness but it does not reveal the root-cause....<CLICK> you will need unified visibility of application performance in the context of infrastructure health --- in real time --- to isolate to an actionable root-cause....
27. I added this just to save us some time...if we had a database bottleneck in an SQL server, <CLICK> ...we could drill down directly to the root-cause, which in this case is a blocker query in the SQL server database... ...the point is, whatever stakeholder you are in the organization--- from a .NET developer to a server admin to a service owner--- you can get personalized dashboards, reports and alarm windows that are situationally aware of what’s happening in real time... ...and automatically diagnose where action needs to be taken and why.
28. Our embedded domain knowledge is useful in other ways besides automated diagnosis too....since we collect measurements at every layer of every component and automatically baseline every measurement, we provide a number of out-of-the-box and customizable reports that allow various stakeholders to optimize performance over time... ...this includes how trends associated with digital business services impact the physical and virtual infrastructure...so it’s very useful for getting end-to-end service baselines that can provide actionable insights to help you right-size the supporting ecosystems....
29. Other capabilities of eG Enterprise are also made available to the .NET environment as well, including built-in knowledge sharing, performance-oriented configuration tracking, the ability to create self-healing correlation and autonomous operation.... ...and of course personalized dashboards and reports.
30. ...the Universal Agent enables simple data collection that provides flexibility in design...agent-based or agentless, active or passive....all with a single data collector that can be auto-deployed easily...
31. ...eG Enterprise also provides both a <CLICK> command line and API interface for integration with tool chains, chat or ITSM tools.... ...and the ability <CLICK> for Admins and Developers to add additional tests and measurements or configure monitoring for custom applications... ...which lets the <CLICK> .NET environment be seamlessly integrated into a unified monitoring environment.
32. ....everything we’ve discussed today can be deployed on your premises or using eG On-Tap, which is our cloud or SaaS-based option... ...this allows rapid return on investment, easy installation and the ability to exploit additional capabilities over time and fully leverage your monitoring investment as you build in-house capabilities around the platform
33. So, this is what we mean when we say “full-stack” .NET monitoring.... ...monitoring every layer...user experience, business transactions, .NET applications and infrastructure performance...and automatically isolating the specific layer of the specific component is the cause of the problem
34. ...the benefits to IT and the business are clear... Total performance visibility of the user experience and automatic diagnosis ... ...this results in reduced downtime, leads to proactive management and rapid mean time to restore...and can help optimize the service infrastructure.... ...and of course having a unified view of performance for every service team member can accelerate IT transformation as well.
35. Let’s see what other questions we have.... SEED QUESTIONS What is your differentiator compared to other .NET monitoring tools in the market? Well, the key difference between eG Enterprise and most other solutions is that we’re not just a .NET code monitoring tool...the ability to provide visibility of both .NET and non-.NET components...and automatically isolate performance issues from end user experience, to code, to infrastructure is pretty unique...and it can come at a fraction of the cost of many other APM tools such as AppDynamics and Dynatrace... Does eG Enterprise support monitoring cloud environments? Yes, eG Enterprise offers native support to monitor applications and infrastructures on public, private and hybrid cloud environments... ...for example there are out-of-the-box monitoring metrics and dashboard for AWS and Microsoft Azure platforms. Can you monitor APIs and 3rd party calls? YES...usually this is demonstrated when a customer’s interested in seeing our APM capabilities....and the embedded analytics will automatically isolate to an issue with an API or 3rd party call just like we do with other service components....measurements associated with APIs and 3rd party calls are also auto-baselined like all measurements we collect.... Can we get a free trial? YES...you can request a free trial right from our web site....I would say that in order to get a full view of the monitor’s capability that it would be to your advantage to identify a service....even if it’s SharePoint or something like that...the
36. Just a reminder everyone will get a copy of the slides and a link to the recorded webinar from today’s session