(The Hosting News) – vCenter Operations Manager (vCOPS) is part of VMware’s vCloud suite. Here at Edge we use vCOPS for a number of different management tasks in our cloud hosting environment. But in this blog post I want to discuss how we use vCOPS for monitoring and troubleshooting. There are two functions of vCOPS that help us pinpoint performance issues- infrastructure navigation and proactive smart alerts.
Infrastructure navigation maps dependences between different virtual machines (VMs). Essentially vCOPS discovers and catalogues application dependencies based on the services that are running on the VMs and the traffic flowing between them. vCOPS then visualizes the dependencies in the vCOPS dashboard. Understanding application dependencies helps us troubleshoot problems quickly by pinpointing the root cause.
Take a standard n-tier application- a database cluster, an application server and three front-end web servers. vCOPS’s infrastructure navigation features allow our team to visually view the entire n-tier environment and the application dependencies between them. In our imaginary scenario lets say this environment has a failure, which causes multiple VM’s to alarm. The support team is able to see that the web servers are dependent on the application server, which in turn is dependent on the database cluster. When a VM alarms, vCOPS lets the support team quickly deduce that the issue is at the database layer because it is upstream from the application and web servers. Without vCOPS the support team would have to manually discover the application dependencies, which could be a lengthy process at a time where every second counts. vCOPS eliminates that manual discovery process; greatly speeding up the time to resolution of the issue.
The second function of vCOPS that I would like to discuss is proactive smart alerts. vCOPS maintains a baseline of performance statistics for each VM that it uses to determine alerts. Whereas traditional monitoring tools use static thresholds to trigger alerts, vCOPS dynamically sets individual thresholds for each VM based on the baseline for each VM.
So how does this work in the real world? Say there is a VM, which for the past three months has been hovering around 50% processor utilization. Today, however, this VM’s processor utilization starts creeping up due to a run-away process. vCOPS determines that this is a deviation from the base line and sends out a warning at 60% and then an alert at 65% (approximate dynamic threshold settings). This allows our team to investigate and address the issue before the run-away process has consumed 100% of the processor cycles. This dynamic monitoring of individual VMs gives Edge’s support team a heads up before a problem becomes critical.
vCOPSs is just one of Edge’s tools for monitoring customer environments. But the application dependency maps and the dynamic alerting have become integral components of our support technician’s toolbox. By maintaining and visualizing dynamic maps of application dependencies and providing early alerting based on deviations from performance baselines; vCOPS helps our customers Stay Up and Sleep More.