The nightmare installation story. No IT conference happy hour is complete without one. Funny stuff sometimes–when it’s safely in the past. But smooth implementations make for boring stories. Here’s how to make installing a crucial network monitoring system a boring success story.
You’ve finally made the business case to install a new network management tool that will monitor all your key devices and systems in real-time.
Perhaps you need it to safely ramp up virtualization. Or you need to protect an expanding infrastructure against increasing risks. Or you’ve simply realized your patchwork of separate monitoring tools is letting far too many slowdowns, outages and waste fall through the cracks.
Whatever motivates you to finally pull the trigger, you need this new system to give you long-term value with minimum disruption during its deployment. Accomplish both goals by doing these four things:
Smooth installations begin before the sale. Work with potential providers to establish in advance how much support you should expect from the vendor during installation.
Start by asking yourself two basic questions:
Do we have the internal expertise to install and configure the hardware/software ourselves?
This isn’t necessarily a question of whether your staff is good or bad at their jobs. Even experienced systems engineers can have trouble with monitoring technology they haven’t encountered before.
If you answered “yes” to the first question, do you want to devote your internal experts’ time to installing network monitoring system tools?
If you do the entire installation internally, what WON’T be getting done during that time? If you decide your key employees could be serving your users and your bottom line better by staying at their usual tasks, be sure the vendor can offer the support you’ll need.
The fastest way to get an installation off the ground quickly is to have a complete inventory of the virtual and physical devices you need to monitor. This isn’t always as easy as it looks.
For Netreo customers who aren’t sure which devices should be monitored, our service engineers (SEs) lead them through a series of questions, often starting with, “What’s the end-state you want?”
Say your desired end-state is 99.99% uptime. An SE might follow-up with this: What, exactly, needs to be up 99.99% of the time?
Specific physical hardware?
All of the above?
The answers to these questions will yield at least an initial inventory.
For each device in your inventory, do any necessary configurations so that the monitoring software can find it and talk with it. You may need to configure SNMP on your networking gear, or a WMI protocol for Windows systems.
Some monitoring tools, such as Netreo’s OmniCenter, include an “auto discovery” function that will help you create an inventory. Be ready to configure the devices identified, however, so the new tool can poll each device.
Your goal in this first phase is to prepare your network so the monitoring tool can fully poll each device to create a baseline of thresholds across your environment. The thresholds can be adjusted over time.
Just about every IT pro who does onsite installations has a version of this story: He arrives at the customer’s IT facility, introduces himself, and says he’s ready to begin. The response: a blank stare. An apology. Someone rushes off to find someone else, who does the same…
An hour later, he’s watching while two or three employees argue over which rack he should install your server in, which work station he can use to configure and license it, and so on.
From a start like that, he can usually assume this won’t be the fastest installation of his career. You’d better hope he’s not charging by the hour.
It’s up to the vendor to communicate with you about hardware installations like these, to make sure their people are expected and any preparations have been made. Even when that happens, however, the process can break down if you don’t designate a specific point person to assist.
Choose someone who can handle any access requirements, such as firewall or other security issues. This person should be responsible for confirming that every device you want monitored is configured to connect with the monitoring software.
Outsourced IT functions and operations are a common roadblock to monitoring system installations, mostly in larger organizations.
For example, a network monitoring appliance may need to reside in a specific section of your network, set aside just for management appliances. If this part of your network is handled externally, you need to be aware of this and get proper access to it before the installation date.
Getting past this additional layer of bureaucracy can take time, especially if you’re working with a remote facility in a different time zone. Perhaps the people in the data center don’t have the authority to grant access. Or at least, not quickly.
In short, it’s important to have access established upfront.
When IT folks share installation horror stories, you might laugh and commiserate with them. But don’t let them keep you from implementing solutions that could become critical advantages for your organization
When it comes to network monitoring tools, what should frighten you most is the cost of NOT having a quality system installed and properly configured.
Do you know what it costs your company when you lose a Web server, router, application server, etc., and can’t identify, troubleshoot, and fix the source of the outage quickly?
Monitoring tools are essential—even if you outsource all of your infrastructure. You still need visibility from an independent source into how your applications are performing and how your users are doing.
And by following the steps above, you can have the visibility you need without becoming one of those installation nightmares. You’ll have one less juicy story to tell your peers, but that’s a good problem to have.