Managing events and configurations
What's different about monitoring the virtualized world?
When managing cloud systems (or private cloud virtual systems), there are two dimensions that need to be monitored and correlated:
- Your standard monitoring data - CPU usage, disk usage, memory, processes running, application responing OK, etc.
- The configuration of your managed systems.
Today, the monitoring data is well understood. There are many open source scripts available that can gather just about any metric that you want. And if one isn't already available, a systems administrator can usually create one easily using the scripting language of his choice (Perl, Ruby, shell, Python, etc.). In our view, this level of monitoring is a commodity.
The second item - configuration - is the new wrinkle. The old model is to either manually load the monitored configuration into your tool or intiate an auto-discovery process on startup. However in the cloud, configurations can change at any time (ie, EC2 instances starting and stopping), so how can you ensure that you are monitoring your entire environment?
An additional aspect is tracking the configuration attributes or meta-data in your virtual environment. So rather than just tracking your known host names based on your company's standards, you need to track, in Amazon EC2's case, ami-id, instance-id, security groups, private and public DNS names. This problem is multiplied if you are using services from multiple cloud vendors, each with different meta-data. When looking at service affecting alerts, your operations staff will need access to this meta-data AND the monitoring data to be effective and efficient.
This is one of the problems Tap In is trying to address. If you look at our event management article in the Resources section, you will see that our event architecture merges configuration and monitoring status information into a single event which can be presented to operators, reporting systems, or other automated processes.
What you be seeing in the near future from Tap In Systems is an extension of this concept where any configuration can be encapsulated into the event stream.
