As if by magic ... the dashboard appeared: Building useful Business Service Capacity reports

摘要

1. Collecting data is not too difficult, neither is creating a pretty dashboard. However, selecting the right data, and the right approach to consolidating that into useful information about where to focus your attention, turns out to be a lot more challenging than it first appears. 2. The argument for aggregating the mass of individual component metrics into a set of more manageable service-level dashboards seems self-evident. However, when you have a mapping where a Business Service is comprised of several Applications, each of which is delivered by several tiers, each of which is hosted on several server/VM instances, how should you propagate a warning at the component-level up to the top service-level dashboard? If the primary database goes red, surely the service should also go red, but if it is just one of a cluster of web-servers, the service probably isn't greatly compromised (and even that can only be done once the rules that govern when a server goes red have been defined).