Monitoring systems can help consultants keep an eye on the health of their clients' servers and networks. Learn what steps to take when monitoring critical client systems.
Monitoring critical client servers and networks is a fundamental step in gauging systems' health, yet many consultants (judging at least from the number of clients requesting that my office replace their current provider) miss this important step. Only proactive monitoring and effective performance alerting enables peripatetic consultants to properly support clients.
What to monitor?
Monitoring systems are invaluable. When properly selected, implemented, and maintained, good monitoring platforms help even small consultants keep tabs on client systems' health.
You should look for monitoring and reporting agents that are capable of all the following:
- Tracking backup operations.
- Auditing intrusion attempts.
- Assessing disk health.
- Monitoring critical services.
- Generating real-time alerts.
Business process changes
Unfortunately, the mere acts of selecting and implementing a monitoring platform doesn't guarantee an IT consultant's success. Like most anything else, planning, thought, and follow up is required.
Here are steps consultancies should adopt when monitoring critical client systems:
- Create a list of all monitored systems.
- Assign one engineer to be responsible for monitoring the list each morning. This engineer doesn't have to be responsible for fixing the failures he or she identifies -- just getting the trouble ticket created.
- Create a checklist that tracks all the steps that must be completed when a new monitoring client is brought onboard. Each client should likely have a centrally-accessible checklist that lists which consultancy staff member installs the monitoring and performance agent, which servers are monitored, which traps are set, when the monitoring settings are reviewed (such as quarterly or whenever software, application, or service changes are made), where the alerts are sent, how the alerts are sent (via email, text message, phone call, etc.), which engineer responds to service-affecting outages, how the client is billed, etc.
- Create a new line item within your firm's accounting software to cover the costs of the monitoring platform's license and the daily review of the client's monitoring reports.
- Ensure the client is billed. It's very easy to forget to charge a client for monitoring. Ensure the new accounting line item is charged against the client's invoice by adding this step to the client monitoring checklist, if necessary.
Monitoring method is important
In the past, too many consultants may have relied upon haphazard monitoring efforts. Maybe Small Business Server performance and monitoring reports were configured for some clients, but not others. Maybe reporting features were enabled on an APC management console. Maybe reporting was enabled on some client SonicWALL routers, but not others.
To properly monitor systems, a complete plan (i.e., one that provides consistent, reliable, and actionable reporting across a variety of clients) is required. That's why I advocate implementing a solution from GFI, HyBlue, Quest Software, or Zenith Infotech.
The problem? There are several. Monitoring platforms aren't foolproof. They require their own expertise to master. Clients don't automatically see the value or want to pay for monitoring services.
A real-world example
Monitoring platforms aren't foolproof, and they require expertise to master. However, I find the biggest challenge of these systems is that clients don't automatically see the value or want to pay for monitoring services. When I provide clients with real-world examples of how monitoring platforms have been useful, it usually helps them overcome their objections. Here's one straight from the non-monitored client call logs.
A client calls. We've never done business with this firm before. They're in a panic. It's 4:00 in the afternoon on a Friday. Nothing's working -- no email, no Internet, no printers, no application access. Manufacturing machines have stopped. Orders can't be fulfilled. The company's losing $3,000/hour. Worse, the accountant can't access software required to access payroll records and cannot print employee paychecks, which contracts require be provided within 45 minutes.
The solution? We have to scramble an engineer like an F-18 fighter jet. These are the service calls for which customers would vote, in a heartbeat, "yes" on a state referendum enabling our engineers to respond Code 3, lights, and siren the whole way.
Sales losses, business interruptions, resultant employee discord, and emergency (non-contract client same-day) service fees add up. Worst of all, in this example, the entire debacle could have been avoided, but no capable IT consultant was monitoring this client's servers.
What caused the trouble? An employee picked up a virus infection on Monday. The employee began sending large amounts of email directly from Outlook. The Exchange store grew to consume the entire Small Business Server's hard drive, leaving the OS no hard disk space to manage Windows operations, page file processing, and other critical tasks. The server tanked.
However, its log files were essentially begging for help all week, up to and including the point on Thursday when the Exchange store failed. Disk consumption trends alone would have alerted our engineers, likely Monday evening. If this had been a monitoring client, we would have responded to the issue Tuesday, suspended the single user's account remotely, cleaned her system of the infection locally, and brought her back online. No widespread outages, no lost sales, no employee discord -- everyone else would have kept working.
What are your monitoring methods?
One of TechRepublic's most powerful features is the way in which its members learn from one another. What effective and reliable methods have you implemented to properly monitor critical client systems? Join the discussion.