(“In Case You Missed It Monday” is my chance to showcase something that I wrote and published in another venue, but is still relevant. This week’s post originally appeared on APMDigest)
To help your organization increase data center efficiency and get the most benefit out of your monitoring solutions, here are the remaining universal monitoring crimes and what you can do about them:
Start with Universal Monitoring Crimes and What to Do About Them – Part 1
4. FLAPPING OR SAWTOOTHING ALERTS
When an alert repeatedly triggers (a device that keeps rebooting itself or processes keep deleting/creating temporary page files so that one moment it’s over threshold, the next it’s below, for example), that condition is known as flapping or sawtoothing.
What to do about it: These types of alerts have several possible resolutions based on what is supported by your monitoring solution and which best fits the specific situation:
■ GOOD: Suppress events within a window. Ignoring duplicated events within a certain period of time is often all you need to avoid meaningless duplicates.
■ ALSO GOOD: As mentioned previously, add a time delay to allow for self-resolution, avoid false-positives, and eliminate other potential issues that don’t necessarily require a remediation response.
■ BETTER: Leverage “Reset” logic. Wait for a reset event before triggering a new alert of the same kind. Avoid making the reset logic merely the reverse of the trigger (if the alert is > 90%, the reset might be 90% for 15 minutes, but it won’t reset until it’s
■ BEST: Two-way communication with a ticket or alert management system. This is where the monitoring system communicates with the ticket and/or alert tracking system, so you can never cut the same alert for the same device until a human has actively corrected the original problem and closed the ticket.
5. NO LAB, TEST, OR QA ENVIRONMENTS FOR YOUR MONITORING SYSTEM
If your monitoring system is watching and alerting on mission-critical systems within the enterprise, then it is mission critical itself. But despite the fact that many organizations set up a proof-of-concept environment when evaluating monitoring solutions, once the production system is selected and rolled out, they fail to have any type of lab, test, or QA system that is maintained on an ongoing basis to help ensure the system is maintained.
What to do about it: Duh. Implement test, dev, and/or QA installations that serve to ensure your monitoring system has the oversight necessary for a mission-critical application.
■ TEST: An (often temporary) environment where patches and upgrades can be tested before attempting them in production.
■ DEV: An environment that mirrors production in terms of software, but where monitors for new equipment, applications, reports, or alerts can be set up and tested before rolling those solutions to production. And as mentioned earlier, this is the perfect place to also monitor your production monitoring environment.
■ QA: An environment that mirrors the previous version of production, so that if issues are found in production, they can be double-checked to confirm whether the problem was introduced in the last revision.
Note that I’m not implying you necessarily must have all three, but it’s worth considering the value of at least one. Because “none” is a really bad choice.
FINAL THOUGHTS
The rate of technical change in the data center today is rapidly accelerating and traditional data center systems have undergone considerable evolution in a very short period of time. As complexity continues to grow alongside the expectation that an organization’s IT department should become ever-more “agile” and continue to deliver a quality end-user experience 24/7 (meaning no glitches, outages, application performance problems, etc.), it’s important that IT professionals give monitoring the priority it deserves as a foundational IT discipline.
By understanding and addressing these top universal monitoring crimes, you can ensure your organization receives the benefit of sophisticated, tuned monitoring systems while also enabling a more proactive data center strategy now and in the future.