Being it IT, devops, or no-ops, operations are a critical part of every fairly sized system with real, expressed or not, SLAs. John E. Vincent’s post is an interesting look at what he feels is missing to make system operations more managable:
What I feel like we’re missing is a way to express those relationships and then trigger on them all the way up and down the chain as needed. We’re starting to get into graph territory here.
We must we be able to express and act on changes at the micro level (I changed a config, I must restart nginx) and even at the intranode level (something changed in my app tier, need to tell my load balancer) but now we need a way handle it at that macro level. Not only do we need a way to handle it but we must also be able to calculate what is impacted by that change.
Original title and link: How Can Graphs Apply to IT Operations ( ©myNoSQL)