Zenprise for BlackBerry and Exchange has everything a company needs to monitor and troubleshoot mobile e-mail delivery-as long as that environment comprises BlackBerry smart phones at the edge and Microsoft Exchange and BlackBerry Enterprise Server at the core. Despite this somewhat limited scope of supported environments, Zenprise for BlackBerry and Enterprise hits the mark with an outstanding combination of analysis capabilities and troubleshooting recommendations, earning an eWEEK Analyst’s Choice recommendation.
Zenprise shines because it aggregates its forensic data from numerous networks and back-end applications to identify e-mail delivery problems anywhere in the chain. The product will monitor the health of a company’s Microsoft Exchange and BlackBerry Enterprise Server instances, detect problems in both the enterprise network (including the local Active Directory and DNS servers) and the ISP’s cellular network, and find problems that may exist in Research In Motion’s own mail delivery network. Slowdowns can occur at any step along this complex path, and Zenprise provides excellent insight into each of these elements.
While I’d like to see Zenprise expand its support to include Windows Mobile-based devices using Microsoft Exchange ActiveSync for management and mail delivery, right now that is only on the company’s road map. Given that Zenprise got its start as an Exchange monitoring solution (Zenprise for Exchange), I have hope that support will extend to Microsoft-based devices sooner rather than later.
For this review, I tested a late beta version of Zenprise 4.0 for BlackBerry, which will be available in September. Version 4.0 adds support for 64-bit hardware and operating systems, which allows Zenprise to scale support for a greater number of managed servers (up to 120 AD, MS SQL, Exchange or BES servers with a single Zenprise iteration). Version 4.0 also improves Zenprise support for newer Exchange 2007 e-mail environments.
Pricing for Zenprise for BlackBerry is based on the number of devices supported through the system, so all the centralized elements of the system are rolled into that price-including the centralized Zenprise server software, the reporting engine (which is based on Microsoft SQL Reporting Services), and agents for installation on all the servers Zenprise will manage. For 1,000 users, Zenprise for BlackBerry and Exchange costs about $40 per user.
To kick off my tests, I ran a discovery of my network to find all the components of the mobile mail system in order to add them to the Zenprise database (which runs on MS SQL 2005). The detection discovers all the server nodes for BES, Exchange and the Active Directory infrastructure-plus any BlackBerry devices that are currently managed by BES-and begins monitoring many of them immediately for signs of trouble by collecting standard Windows instrumentation.
However, to collect the BlackBerry-specific device and network information from the BES server, Zenprise requires an agent be installed on all BES instances, which administrators can push out directly from the Zenprise Console once the initial discovery completes. Administrators can also install agents on Exchange and Active Directory components of the infrastructure. The agent allows greater monitoring efficiency and can limit transmissions to known errors. It also can store and forward logs in case of network outage, and allows the administrator to perform real-time diagnostic tests for various services.
Zenprise administrators should run the discovery process frequently to recognize changes made within the BES server, as newly activated BlackBerry devices will not be recognized until the discovery occurs. Discovery events can be run manually or automated periodically.
Achieving High Levels of Accuracy
With the Zenprise database populated, Zenprise begins analyzing symptoms of potential trouble throughout the network. As more data gets collected, Zenprise correlates symptoms together to establish a confidence level of the accuracy of its predictions. To achieve high levels of accuracy, Zenprise uses a technology called ZenPro, which works its way through thousands of complex decision trees to take the discovered symptoms and turn around the underlying causes.
The decision trees were created by Zenprise engineers to look at every scenario imaginable for a given symptom, running through if-then scenarios to see whether collected data supports one theory over another. As more data comes to light to support certain hypotheses, ZenPro first makes an educated guess as to what the problem is (a Probable Cause), then promotes the problem to a Root Cause when the system has full confidence in the diagnosis. ZenPro can take 7,000 detected events representing 400 distinct symptoms and synthesize that down to about 16 problems to resolve, a world-class timesaver if I’ve ever seen one.
Zenprise administrators can see this intelligence at work in the Diagnose and Resolve tab of the Zenprise Console-a Windows-based application that can be utilized for anything from help desk-based remediation and testing of end users’ smart phones to helping an engineer troubleshoot the MAPI connections between the Exchange and BES environments. Once a Root Cause is determined, Zenprise presents a graphic depiction of the workflow (with links to supporting documentation including Microsoft or RIM knowledgebase articles and more detailed remediation advice) to help resolve the problem.
Help desk administrators can only see the User Dashboard, which provides tools that help troubleshoot a device and application performance issues. Once a user reports a problem, help desk administrators can quickly find the right device (by user’s name, phone number, e-mail address, carrier or device type) to see the device status and Exchange characteristics.
Help desk administrators can see a device’s available and maximum memory, associated BES and Exchange servers, network type, and device IMEI (International Mobile Equipment Identity) number. Zenprise also displays the signal strength and battery level for the device as it was last reported by the BES server, although this data may be out of date for use in troubleshooting (depending on when BES last collected the information).
Real-Time Diagnostic Tests
Administrators can also run real-time diagnostic tests-they can ping the device to see if it is currently on the network or to test roundtrip delivery time for a test e-mail message to look for delivery latency or failures. On a separate tab, the User Dashboard shows information relevant to the user’s Exchange instance, showing whether maximum mailbox quotas have been exceeded or whether the password has expired.
Other diagnostic screens, like the Zenprise Dashboard (which breaks down collected and identified problems by managed server), Diagnose & Resolve Dashboard (which presents the corrective workflow for found Causes) and Performance Charts, are only available to administrators in the Zenprise Administrator or Zenprise Infrastructure roles. While both roles can see all data in the Zenprise console, only Zenprise Administrators can change settings or notification rules.
I used Zenprise to troubleshoot a few issues common to BlackBerry activation and usage. For instance, Zenprise diagnosed an activation problem with a device switching from RIM’s hosted BlackBerry Internet Service to the enterprise-oriented BES environment, correctly recommending that the device be wiped before activation and giving the proper steps to carry out that recommendation given the model of the phone. Zenprise also correctly identified when BlackBerry activation messages have been captured by Outlook’s Junk Filter, thereby stalling the activation process.
In addition, Zenprise quickly diagnosed when mail delivery stalled due to an overflowing mailbox or when a device disappeared from the network when messages were queued for delivery.
BlackBerry users can be organized into VIP groups (Zenprise currently supports a maximum of five VIP groups), clustering together the most important users in the company so they can be monitored separately with a single console view. For instance, I could organize all my C-level executives into a group to ensure rapid response time if one of their devices experiences a slowdown.
However, I was disappointed to find that I could not set up email notifications particular to one VIP group to my satisfaction. While I could set up notifications to alert administrators when many different types of events occur, I could not localize these notifications to a single VIP group. In other words, if an IT executive wants to know when the members of the most critical VIP group (composed of C-level executives, for instance) experience an outage of more than 15 minutes, the IT executive must accept similar notifications for the members of all other VIP groups as well. While this point may be nitpicky, the lack of such granularity undermines some of the value of the VIP classification system.
Diagnosing Problems Within BES, Exchange Environments
On the infrastructure side, Zenprise provides invaluable assistance in diagnosing problems within or between the BES and Exchange environments. Zenprise can detect problems like hung threads, MAPI errors or high disk usage that could bottleneck mail delivery. The service monitor also helped quickly identify big problems, within a minute notifying administrators when I killed a BES process or disabled the network adapter on the Exchange server.
Zenprise offers outstanding reporting capabilities and the ability to create highly customized performance charts for the BES and Exchange environments. Among the cascade of available reports, I could generate usage trending reports, root cause analyses and system configurations, or I could examine the mean time to resolutions to evaluate performance against a service-level agreement.
For performance charts, I could examine CPU, disk or RAM performance across a selection of servers (over an hour, day, week or month), or I could drill down into a vast array of application-specific metrics, anything from message latency through the BES system to the recent performance of the OWA Web mail interface.
However, administrators should know that Zenprise could have a management and reporting issue at the largest scale. Each Zenprise server is managed separately via its own iteration of the Zenprise Console, and the back-end Zenprise databases will likewise be distinct.
While global companies could isolate Zenprise server instances and management responsibilities according to region, I’d like to see Zenprise develop a manager of managers console to aggregate Zenprise management and reporting for the largest companies that want to keep ultimate control under one roof.
eWEEK Labs Senior Technical Analyst Andrew Garcia can be reached at [email protected]