Tips for Deploying a Performance Monitoring Platform
Deploying and integrating a new performance monitoring platform into your infrastructure operations may seem challenging, but it doesn’t have to be.
We recently sat down with SevOne’s Brandon Butler – a Senior Solutions Architect – who has a few tips you can use to make the process easier on everyone involved.
Each day, Butler – and the team of professional service engineers – has the opportunity to work directly with highly skilled customers to ensure that everything is running smoothly.
Sure, some days the team has to put out fires. But most of the time, they’re advocating for our customers and checking in with them to see that their needs are met. The professional services team also helps bring customers up to speed with the newest appliances and most up-to-date versions of the SevOne monitoring platform.
SevOne: What is the first thing new customers should do before deploying a performance monitoring platform?
Brandon Butler: We have two common customer categories – a proactive group of customers that wants to use performance monitoring for capacity planning and forecasting, and another, more reactive segment, that seeks out a performance monitoring platform for alerting and alarming. Our largest customers use performance monitoring for both of these reasons.
All of our customers need to make wise financial decisions and rely on our platform to make these decisions. Most of our customers have previously used performance monitoring – they either have a good idea of how they want to do it or they’ve done it before.
The first thing we do is ask why and how they’ll use performance monitoring. Then, we’ll make equivalent reports and dashboards to what they’re used to using. We’ll take it one step further by adding additional value, making the data much more granular and specific.
S1: Are there any other basics customers should be aware of before launching a performance monitoring platform?
BB: Choose your naming conventions wisely. The performance monitoring platform is usually quickly adopted. And you’re dealing with big databases here, so the better you can sort devices and groups, the more you’ll get out of the product. Having a uniform naming convention makes it easier to apply filters and apply text-based rules. Consider naming your devices and groups with code. Naming a device by city, state, type of device and the role it plays in the network might look like this: PHL PA SW1 CUS. This, for example, is a customer switch located in Philadelphia, Pennsylvania.
S1: Should enterprises and service provider organizations worry about monitoring capacity planning yet?
BB: We’ve never had a customer say they wish they had gotten a smaller amount of monitoring capacity. Get the most capacity you can afford when deploying a performance monitoring platform. We’ve had tickets come in from customers who were no longer receiving reports. Coworkers had turned off others’ flows so they could get their own because the box had reached capacity. Monitoring capacity fills up quickly, so consider investing more heavily early on.
S1: How critical is it for organizations to have a monitoring backup in place from day one?
You may not think you need backup, but deploying a hot standby solution – a duplicate database – could prove critical someday. Many organizations don’t consider a backup until it’s too late. The upfront price tag is marginal compared to what it would cost if you lost a year’s worth of polling data.
A hot standby solution replicates any configuration and polling data. When a change is made on the main database, it replicates within 15 minutes on the standby solution.
You don’t want to risk losing your data for a few reasons. First, for capacity planning, you need historical data. And if you need to predict activity for the next six months, you can’t make an informed decision if you’re missing last year’s data. You may also be able to avoid SLA penalties, since there won’t likely be any gaps in coverage. You may currently incur penalties for not being able to recover data within one hour, two hours, maybe five hours after an outage.
For more best practices, download our free whitepaper on 6 Steps to an Effective Performance Monitoring Strategy.