Your uninterruptible power supply (UPS) system is your data center’s first line of defense against power variations—spikes, dips and so on. Because thousands of dollars (or far more) of IT equipment is depending on a clean power supply, you can’t afford to wait for problems to arise in your UPS system before taking action: preventive maintenance is critical to uptime and to protecting sensitive electronics.
According to Eaton Corporation, some two-thirds of downtime events are preventable, with insufficient maintenance being one of the main causes. You may not be able to do much about malicious attacks, failures on the part of service providers and unforeseeable equipment failures, but you can take steps to ensure that many potential downtime events are thwarted. Preventive maintenance spans the entire range of systems in the data center, and each system has its own needs. In the case of the UPS system, here are several tips (in no particular order) to help ensure that power fluctuations don’t threaten your IT equipment.
Tips for Preventive UPS Maintenance
- Put safety first. Life and limb should trump dollars every time. When you’re dealing with electrical power, you’re always one small blunder away from serious injury or death. So when dealing with UPS (or any electrical system in the data center), make sure that safety is a top priority: that includes observing manufacturer recommendations, paying attention to the peculiar implementation details in the facility and following standard safety guidelines. If you’re not sure about some aspect of your UPS system or how to maintain or service it, call a professional. And even if you do know your UPS system inside and out, getting outside assistance can still be warranted—if for no other reason than it involves potentially cooler heads that aren’t dogged by the same pressures.
- Schedule maintenance—and stick with it. Preventive maintenance shouldn’t be something that you’ll just “get around to,” particularly given the potential costs of downtime. For your UPS—as with other data center systems—you should schedule regular maintenance activities (annual, semiannual or whatever the time frame) and stick with that schedule. That includes keeping a written (on paper or electronic) record listing upcoming maintenance activities and whether/when past maintenance was performed.
- Keep detailed records. In addition to scheduling maintenance, you should also keep records of the kinds of maintenance performed (for instance, cleaning, repair or replacement of certain components) and the condition of the equipment during inspection. Keeping track of costs can also be beneficial when you need to show the C-suite that a few dollars in maintenance costs beats thousands or millions in downtime costs every time. A checklist of tasks, such as inspecting batteries for corrosion, looking for excessive torque on connecting leads and so on, helps maintain an orderly approach. All of this documentation can help when planning for equipment replacement or unscheduled repair and when troubleshooting the UPS. In addition to keeping records, be sure to keep them consistently in an accessible and well-known location.
- Perform regular inspections. Much of the above can apply to almost any part of the data center: enforcing safety, scheduling maintenance and keeping good records are all excellent practices regardless of the data center context. For UPSs in particular, however, some tasks can and should be performed regularly by staff (who should become familiar with at least the basics of UPS operation). A few important UPS maintenance tasks include the following:
- Visually inspect of the area around UPS and battery (or other energy-storage) equipment for obstructions and proper cooling.
- Ensure no operating abnormalities or warnings have registered on the UPS panel, such as an overload or a battery near discharge.
- Look over batteries for signs of corrosion or other defects.
- Consult manufacturer documentation for guidelines and recommendations for your specific equipment. You should perform maintenance (or hire professionals to do so) as often as the manufacturer recommends at a minimum—in some cases, though, more often may be better.
- Recognize that UPS components will fail. This may seem obvious: anything with a finite probability of failure will fail eventually.
- Know whom to call when you need service or unscheduled maintenance. During daily or weekly inspections, problems can arise that may not be able to wait until the next scheduled maintenance. In these cases, knowing whom to call can save a lot of stress. That means you must identify solid service providers that will be available when you need them (i.e., at odd hours). The provider may be the same as your regular provider or not—but keep all the numbers for maintenance and repair in a readily accessible location. (And if you have your well-kept maintenance records in the same place, you will be able to supply helpful information when the provider arrives, potentially saving lots of time and money in service fees.)
- Assign tasks. “Weren’t you supposed to check that last week?” “No, I thought you were.” Avoid this mess: ensure that the appropriate personnel know their responsibilities when it comes to UPS maintenance. Who checks the equipment weekly? Who calls to schedule (or perhaps adjust the schedule for) annual maintenance with the service provider? The particular tasks may vary, but make sure you know who is responsible for what when it comes to your UPS system.
Many of the particulars of UPS maintenance are best left to those who know UPSs in general and, more importantly, your UPS implementation in particular. Again, safety is critical: voltages inside a UPS system are at lethal levels, so it’s better to spend a few dollars hiring a professional than to risk lives (and, secondarily, downtime). Preventive maintenance is critical to all aspects of the data center, so many of the tips above can apply broadly, but the UPS system requires particular attention because it is the short-term guarantor of a clean, steady flow of power to your IT equipment. Regular, scheduled maintenance can easily pay for itself by preventing unplanned downtime events thanks to battery or capacitor failure, clogged air filters, welded relays and even outdated firmware. Taking a few organized and deliberate steps now to implement a preventive maintenance program can save many UPS headaches later.
written by Jeff Clark, Reprinted from Data Center Journal