NetIQ Aims To Balance Simplicity, Flexibility

NetIQ has been in the APM game longer than most, but it's being challenged by a huge roster of rival application performance management vendors. And no wonder: This is a growth market, according to Forrester Research, which expects vendors to net almost $2.1 billion by year's end. Seems there's no time to waste on sluggish apps.

One challenge for new and established vendors alike is keeping their APM products easy to use in the face of increasing network complexity, while still providing for customization. NetIQ's AppManager 7, part of the NetIQ AppManager Suite, uses agents to provide monitoring, reporting, analysis, diagnostics, and resolution of common application issues. It mostly succeeds on the ease-of-use front, as we found in our testing. To further enhance the product's appeal, NetIQ offers optional modules geared toward monitoring such technologies as IBM WebSphere, Microsoft SQL Server, Oracle, and VMware ESX Server. We also evaluated NetIQ's Analysis Center, using it to analyze performance data and run complex reports.

INVISIBLE TOUCH

NetIQ advertises a zero-touch capability that lets administrators deploy agents from a central console. For this feature to work, however, the management server needed to be in an Active Directory domain with the client systems. Unfortunately for us, our test bed is distributed, so we had to install our agents manually.

We were most interested in trying out AppManager's ability to monitor end-to-end transactions, and we weren't disappointed. We configured AppManager to monitor a multitiered application, end to end, using a combination of synthetic transactions and an agent that looked at the performance of both the overall system and individual applications, such as Web servers and databases. NetIQ's Knowledge Scripts are the main technology behind AppManager monitoring. Knowledge Scripts are VB or Perl scripts that contain instructions for the agent to gather the data being requested. They ship with a number of preset thresholds to report performance problems and faults. Thresholds can be adjusted to suit your needs, but we found a lot to get started with right out of the box.

AppManager ably reported on specific performance issues related to multiple independent components, including our Web and database servers, and placed discrete issues into the context of a logical application. It also did a very good job identifying specific problems, such as the failure of a process or step of a synthetic transaction. However, putting these alerts into the context of an application failure required manual intervention. Whether this will qualify as onerous depends on how complicated your application infrastructure is.

BUT CAN YOU MAINTAIN IT?

AppManager 7's Service Maps--graphical representations of the network, applications, and services--chart application dependencies. Service Maps help isolate root cause and show the logical flow of application performance.

We built a simple map depicting an application that combined system-, network-, and application-related metrics so that individual alerts, when they occurred, were displayed in the context of the application as a whole. During testing, we saw critical alerts depicted in the service flow so that root causes were clearly identifiable.

The downside is that you'll need to build Service Maps manually. This requires an understanding of all the components of an application and how they relate. And, while NetIQ calls Service Maps "self-maintaining," that assumes there will be no changes to the application infrastructure.

Back in the real world, Service Maps must be manually updated, because large, distributed applications change form over time. In a dynamic environment, this level of maintenance could get tiresome fast. And if you use virtualization, that will definitely make "find the server" a much more difficult game. However, while the virtual machine may move from physical host to physical host, the IP and routing information won't. Further, as performance statistics are represented as a percentage of the total, assigning additional resources to a VM to adjust for performance shouldn't break any of the Maps.

We liked the Analysis Center tool, which enabled us to produce detailed application reports showing the Web response times of a synthetic transaction on a step-by-step basis, as well as health statistics for our database, Web server, and system all in one report. Graphs can be as granular as your raw data permits and can go as far back as you've enabled retention.

THE BOTTOM LINE

In terms of ferreting out information, AppManager's use of agents definitely gives it a leg up on agentless or appliance-based competitors. Organizations will need to weigh whether the depth of information AppManager yields is worth the resources to swap out or install and maintain new software agents on every box.

We also liked the Service Maps, and if AppManager could've built and updated application services on the fly, this product would be a home run. As it is, however, many organizations won't take full advantage of this capability.

Overall, we found AppManager on par with other agent-based APM products we've seen, so if agents aren't a problem for you, consider whether you can commit to the care and feeding needed to make this product sing. NetIQ will likely have a lead over rivals in areas where it supplies application agents; click here for a list of supported agents, which includes VMware ESX Server and BlackBerry Enterprise Server.

While à la carte pricing is available for the various components of the NetIQ AppManager suite, the product also is sold in foundation packages that bundle various consoles, modules, and capabilities. Pricing for foundation modules begins at $995 per server. The foundation bundle that includes all the components tested in this review would cost $1,595 per server.

Check out our Rolling Review kickoff. And click over to our Application Performance Optimization Immersion Center.