Because Microsoft Exchange Server relies on so many separate systems to do its job, troubleshooting can be problematic and require the involvement of any number of IT specialists. I've even been tempted to call the janitor and ask him to reboot the vacuum cleaner, just in case that might help.
As an administrator troubleshooting problems, I'm often frustrated by "noisy" monitoring tools that choke the console with apparently unrelated events, requiring me to figure out which events are causes and which are effects and adding to the downtime. Enter Zenprise, a software tool that detects and resolves problems with Exchange Server 2003 and Exchange 2000 Server. Zenprise identifies DNS servers, domain controllers (DCs), Exchange Server machines, and the other resources that Exchange requires and monitors these resources for baseline performance activity. Using that baseline data, the software formulates performance thresholds and records events when the thresholds are surpassed. Many monitoring tools offer the same capability, but Zenprise goes one step further and uses a series of industry-standard algorithms to draw correlations between multiple events. From this data and ruleset, Zenprise determines a root cause and proposes a course of action.
I tested the Zenprise application on the company's demonstration environment: a network with an Exchange 2003 Server front end and back end and two DCs. The installation process was straightforward; in this small test environment, the process took half an hour. After installation, the program identified servers and resources, which I could elect to manage with or without agents. Some organizations prefer, for security or architectural reasons, to use the agentless mode. I used the agent monitoring in my test.
Zenprise can effectively distill a series of alerts down to their root cause. For instance, when a DNS server becomes unavailable, clients can't reach the Exchange Server, and several other DNS-reliant processes report errors. In the tests that I performed, Zenprise was quite capable of doing the sifting and identifying the root cause of the correlated symptoms. It automatically pursues diagnostics by running Eseutil and querying event viewers for specific correlated data. Using these correlations, Zenprise promotes a suspected cause from "probable" to "root." Zenprise then generates resolution instructions in a flow chart that shows the steps you need to take, as Web Figure 1 shows.
In the situation I just described, a flow chart directs you to check that the affected DNS server can be pinged, confirm that the DNS service is running, then check for corrupted DNS records. The flow chart is easy to follow and is customized with the computer names of your environment—a nice touch.
When I used Zenprise to troubleshoot simulated problems with the Information Store and mailboxes, I appreciated the flow chart. However, at one point in my testing I realized that I was following the wrong resolution path. When I interrupted the process to go back and restart the correct repair process, Zenprise generated an unexpected error. This wasn't a terminal problem, but it took me some time to get the application back on track. If Exchange goes down and users are breathing down my neck, I won't feel like troubleshooting my troubleshooting tool. I spoke with Zenprise personnel who confirmed that the latest release of the software resolves the problem I experienced. According to Zenprise, the company releases new versions of the product on a six- to eight-week cycle; Zenprise subscribers receive free product updates.
When Exchange needs attention, anything that reduces your time to resolution is welcome. Zenprise can help identify problems with Exchange and offer resolution. The problems that I experienced during my testing indicate to me that Zenprise has a version or two to go before it reaches maturity. However, although the price seems a little steep, this application could show value for many Exchange administrators. In addition, the fact that the product is available as a subscription suggests that Zenprise will address bugs regularly.
Editor's note: Version 2.5 of this product was expected to ship as this issue went to press.