These topics are all the rage now and have taken center stage in the virtualization, utility, cloud and next generation datacenter discussions. Great stuff, good things to shoot for, but on an even more fundamental level, anyone who’s attempting automation, autonomics, run book automation, etc. without having a solid implementation of the fundamentals of network, systems, application and service management and monitoring in place is, frankly, F-O-O-L-I-S-H.
You’ve absolutely got to have the right level of visibility into the environment first. When some automated activity takes place, you’ve got to know if it was successful, and the impact it may have had on the bigger end-to-end service or application, system or network environment. The only way for this to happen is to ensure instrumentation and eventing is in place, have a consolidated (network, server, application, service, mainframe, distributed, middleware, transactions, virtual, vmware, cloud, etc.) event view with analytics in place to ensure that operations focus is on the right events and a presentation layer that ties everything together into a business service context.
That’s hard to do. It’s no wonder that the things that are first to go are patch management, operating system or core system (CPU, Memory, Disk/Storage) automation activities. They’re low(er) risk, repetitive and time consuming tasks. What group owns these tools? The systems administration groups for the most part. Do they have an end-to-end business service focus? Do they understand (or care) about the impact on the bigger picture? Do the tools used provide eventing, status and insight into what they’re doing? Are they integrated into the broader IT management and monitoring architecture? Would the operations group know how to respond to a “CPU Provisioning Event” if they even got one? Would they know what system, application or business service it’s impacting? Would they know how to prioritize this event against the hundreds of others they are faced with?
What are your thoughts or experiences here? How are you operationalizing investments into automation, autonomics or run book automation beyond the systems administration groups? Can you get useful eventing from these tools? What’s the level of effort required to instrument an “autonomic lifecycle”? What’s being done to align these “sexy” buzz words into the Business Service Management (BSM) story?
Comments on this entry are closed.