BELLEVUE, Wa – Ed “The Scripting Guy” Wilson and Joe Levy from Microsoft’s Operations Management Suite (OMS) team presented the current state of their product today as part of the 2016 Global Powershell and DevOps Summit. I was not familiar with OMS before today, but the use case it’s trying to fill is a huge pain point in cloud ops: how do you sanely do configuration management across multiple system types, in multiple locations?
From what I understand, OMS is basically a giant job scheduler in the cloud. You install their agent on your servers (doesn’t matter where the servers are – Azure, AWS, on premise, whatever -as long as your networking rules are sufficiently permissive) and set up configuration scripts to run at scheduled times on different categories of systems. This being the new Microsoft and all, the target machines can be Linux as well as Windows. The OMS service has some nice notification capabilities (it will send you an email if it enforces a configuration change on a server) and also integrates with Powershell’s Desired State Configuration (DSC). (Yep – it’s Powershell as a Service. What a world we live in.)
I liked most of what I saw, though I did have a couple concerns. Microsoft may heart Linux now, but Powershell/Powershell Workflows are still the only runbook types allowed on OMS, so if you really want to manage Linux servers this way, you’re going to be using some clunky third-party SSH workarounds. I asked if support for other scripting languages like Bash or Python would be added soon, but there are apparently no plans for this. The source control integration for the scheduling scripts also seems a little weak; it’s apparently Github-only right now, but I would want to play with this before making a judgment one way or the other.
In short, I think OMS already seems like an awesome configuration management choice for people managing big Windows ecosystems, but the product still has a little way to go before people in hybrid or Linux-only environments are likely to get on board.