Archive for January, 2012

Five Keys to Success with APM in Production Environments – Enterprise Scale and Readiness (Part 5 of 5)

By Diego Lomanto (Twitter: diego_lomanto)

Welcome to the concluding post in our series on the Five Keys to Success with APM in Production Environments.  In this series we have been discussing how the Gartner Magic Quadrant provides a great start to implementing an APM solution.  However, maximizing your APM investment in production hinges on critical capabilities that can make or break an implementation.   To refresh your memory, so far we’ve covered:

Now, let’s talk about enterprise scale and readiness

Enterprise Scale and Readiness

If you can check off everything above – you’re in good shape.  But there’s one more factor to success with APM in production.  When you roll out your investment, will it be able to handle the scale of your business today and when it grows?  If you are investing in APM, then you probably have a high-volume of critical transactions to analyze.  Your solution must be able to handle hundreds of millions of transactions per day.  It cannot fail just as transactions sharply rise.  These are precisely the times you need APM!  It goes back to the whole continuous monitoring vs exception-based monitoring argument and ensuring that the solution does not fail at crunch time, or work around limitations by limiting scale of deployment.

You must also be sure that the solution itself can scale easily as your environment grows.  IT is fluid, and APM must be as well.  It’s not just about number of transactions either.  Determine if it’s easy to add tracking on new tiers or is that a whole project unto itself?  Make sure your vendor provides a path to application monitoring expansion that is manageable.

Speaking of manageability, some other key requirements are that the solution should not have a single point of failure, should be able to be remotely configured and should have a high availability.  In addition, whatever method of tracking employed should have low overhead so you that it doesn’t crash once you put it into production.  This is typically a problem for exception-based solutions so keep your eyes on that.

Now, You are Prepared

Ok, now we you have a better handle on the ins and outs of APM in production.  It’s important that you pick a product that not only includes the five dimensions of APM, but also can run effectively in production and provide the right information to you at the right time.  If you look for these requirements during the research phase, you’ll have a much better experience during roll out.  And your solution will provide much more value, both for IT and the enterprise.

Thanks for tuning in during this series.  It’s been great walking through these critical success factors and judging by the traffic it’s helped a few people.  If you have any comments, I would love to hear them in the comment box below, or on twitter @diego_lomanto.

diego_lomanto

January 17, 2012 at 4:08 pm 1 comment

Five Keys to Success with APM in Production Environments – Broad Platform Support (Part 4 of 5)

By Diego Lomanto (Twitter: diego_lomanto)

This is the fourth of a five part series where we explore the critical factors of implementing APM in production environments successfully.  You can find partsone , two and three here.  Please check back next for part five. 

In this series we are discussing how the Gartner Magic Quadrant provides a great start to implementing with APM solution.  However, maximizing your APM investment in production hinges on critical capabilities that can make or break an implementation.  Capabilities that don’t get as much coverage in the media. They are:

Part 4 – Broad Platform Support to Eliminate Blind Spots in your Monitoring Strategy

Your APM solution must be adaptable to support everything you have within your environment.  This includes support not just for application processing tiers, but also databases and middleware, as they are often the cause of performance problems.  A complete APM solution should support a diverse mix of SOA, private and public cloud, middleware, databases, homegrown, legacy and proprietary technology stacks. A solution that will thrive in production should also support applications of unknown design/no access to code.  This provides IT with visibility into 3rd-party applications’ and components’ performance and reduces the time to identify performance issues associated with 3rd party applications and components.

At the heart of broad platform support is the capability to  track each and every transaction instance through its entire life cycle. To enable this, there are multiple techniques to track transactions across virtually any application and environment. Some of these innovative technologies include

Active Context Tracking - ACT technology uses lightweight agents to track cross-tier context and the resource utilization of each transaction across the entire datacenter.

APIs – C, C++, C#, or Java APIs for monitoring transaction flow through proprietary and homegrown components.

Network packet capture – Non-intrusive agents capture all network traffic to identify and measure transactions from the network perspective.

Real-time log parsing – For platforms that are not optimal for instrumentation, real-time log file analysis is used to identify information about transaction flow. The log analysis is seamlessly integrated into the transaction model with no overhead

Passive Context Tracking - When a transaction crosses a component that cannot be monitored (for example, a security appliance that cannot be instrumented and does not have logs), passive context tracking is used to stitch together the flow using an ID common to the different parts of the transaction.

When a transaction enters the system, it is identified by one of the agents and then undergoes classification and analytical processing.  An effective production-ready APM solution continues to track the transaction as it traverses web, application, middleware, and database tiers, while collecting performance and resource consumption metrics at each tier. Even when a transaction makes a call to a tier that isn’t monitored, metrics such as the number of calls and the time spent on that tier are captured.

As we discussed in parts one-three, success in production is contingent upon continuously monitoring as much as possible to derive intelligence right when it’s needed.  The capabilities listed above, out of the box, are important part of the always-on discovery and tracking approach to APM. It is important for the speed of deployment and value provided by the APM solution.  If you have wide gaps in coverage, you simply miss too much and have too many blind spots in your monitoring capabilities.

So, that’s the fourth critical requirement.  Coming up next, and closing out the series, is enterprise scalability.


January 5, 2012 at 8:25 pm 3 comments

 
Subscribe in a reader

OpTier Application Performance Management

OpTier Twitter

Follow

Get every new post delivered to your Inbox.