Data center power outage alerting

May 26, 2009 by
Filed under: Data Center, environmental monitoring, Intelligent Sensors 

monitor power outagesWe have already discussed the importance of monitoring temperature and humidity in the data center. Making sure that the investment your company has made into its networking hardware is well protected is paramount. However, normal temperature fluctuations are not the only potential danger to data room equipment. Power outages are another potentially hazardous event.

You’ve probably already placed UPS devices at the appropriate positions within the data center; which is the first step to protecting your investment. But if the power fails in the middle of the night and your UPS kicks in to keep your vital equipment running, how long will the battery backup keep your servers up and running, or better yet, your air conditioners and HVAC systems?

With the Ravica AC Voltage Detector, you can receive alerts when your systems switch to battery backup. Knowing immediately when your equipment is no longer running on line voltage can mean having the extra time to get a generator up and running or the core power issue resolved before batteries run out of juice and other systems begin to fail.

Temperature can even come back to bite us in this scenario. If your data room isn’t getting the proper cooling and ventilation, because the air conditioners or HVAC systems are no longer operational, you are now not only contending with power failures, but data center hot spots and equipment overheating as well.

The AC Voltage Detector is a simple sensor that works with any of the Ravica SensorProbe models. It detects voltage at 50 VAC to 250 VAC and runs on a open/closed contact switch system. You simply plug the cable into your UPS, like you would with any other piece of hardware. When the sensor detects the presence of line voltage, it reports a NORMAL status. When the power fails and the UPS switches itself on, the sensor automatically detects the change and reports an ALARM status.

Perhaps not every scenario calls for that extra line of defense, but when you consider the cost of data room equipment and the potential cost of that equipment’s failure, some extra measures to ensure its stability just make sense.

~ Jon Mills
Follow me on Twitter

Related posts:

  1. 5 tips for an energy-efficient data center
  2. Higher temperatures keep data centers cool for U.K.’s weather forecaster

Tags: , , , ,

Comments

  • Mike Hamershock

    Since the Ravica BitSight units are single corded for power, what do I plug it into? The same UPS string as the sensors are on? A different UPS string? What happens when the UPS powering the BitSight drops? Then I lose all visability to the sensors. The Bitsight heads really need to have dual corded design so that they can stand on 2 separate UPS strings.

  • http://www.ravica.com Jon Mills

    Hi Mike,

    Really, you can have the BitSight getting power from anywhere you want. It can be the same UPS the sensors are monitoring or a different one all together. If there is an emergency life line string then throw it on that. The point of having a BitSight in place to monitor your UPS is more to know when you are using battery backup vs. line voltage, not when your UPS is failing. Ideally the Ravica systems in place have already notified you that you are running on limited battery life well before the UPS is drained.

    The fact of the matter is, if your BitSight is on an emergency life line string and the UPS supplying its power goes down, then any systems that it would use to alert you (e.g. web servers, mail servers, VoIP servers, etc.) are going to be without power as well. Which would make keeping the BitSight online a moot point.

    But like I said, the real idea here is to be proactive and get the alerts well in advance that say, “Hey, your power went out and you are on battery backup. Time to shut non essential systems down and keep business critical apps running for as long as possible.”

    Hope that offers some additional insight.

  • Mike Hamershock

    Thanks. I’m not too concerned about my servers – all our gear is dual-sourced for power, and each cord is off of a different UPS feed. Going on UPS or even loosing an entire UPS feed “should” have little to no impact due to our design, and so far in practice this has proven true.

    My concern is that I have experienced (more than once) a complete loss of UPS feed when a UPS goes to bypass due to an internal fault and the STS does not fall over to the reserve (secondary) UPS. This is mostly an issue at colos since I don’t control the design – some don’t use an STS and rely on gens for feed when a UPS goes to bypass. When this occurs and the BitSight is powered by that string, I will lose all data from the sensors on that Bitsight. We used to have this issue with Cyclades remote console systems – they’ve addressed the problem by having dual power inputs on their gear now. If I am to rely on Ravica units as a part of my enterprise monitoring system, and if they are to be considered critical systems, they need similar redundancy. I’m just sayin….. ;)

  • http://www.ravica.com/blog/sensorprobes/server-room-temperature-alerts-via-skype/ Server room temperature alerts via Skype : Ravica Blog

    [...] is just another one of the many ways the SecurityProbe can help with event notification. Did the power go out? Is the humidity way too high? Is air not flowing to a part of the server room with known hot [...]

  • http://www.ravica.com/blog/sensorprobes/dry-contact-monitoring-with-ravica-securityprobe-x2060/ Dry contact monitoring with Ravica SecurityProbe-x20/60 : Ravica Blog

    [...] of things. For instance, dry contact solutions are often used with burglar alarms, smoke alarms, UPS, AC units and many [...]