I would like to add one or two more points points to my previous post - An Open Letter to the OpenStack Foundation.
- I used Ceilometer as an example, and only as an example. I am more than sure that there are perfectly valid reasons to have chosen MongoDB instead of MySQL. I have no doubt. But doing so this still introduces more complexity.
- Paul Richards left a comment – and because of his example – I think this just made thoughts even clearer.
"The point is to use the right tool for the job. Don't use pliers when you need to use a wrench".
I completely agree with that. I “stole” his example and used it to further emphasize my point.
Of course every tool has to fit the right job, but to steal your analogy (forgive me), each group thinks that their tool is the right one. And over time - instead of walking around with only a pliers and a wrench you now need a big toolbox with a hammer, a saw, 15 different screwdriver bits, a cutters, duct tape, wire, a wrench, a pliers and so on and so on.
To continue the analogy - I would have like to see the TC either change the bolt - whether now or in the future - that it will become acceptable for both tasks with one tool. Or alternatively create a tool that can be used both as a wrench and as a pliers.
That is the job of the TC. But I personally feel that since the work being done on the overall architecture, a holistic OpenStack Product architecture, is lacking - here is where the biggest potential failure could occur.
I already see it breaking apart - with different HA models, different scaling models for the different components. And it is only going to get worse.