We mostly use it to see stack utilization. Corporate uses it for its break down of teams, who is being charged what, but there are a lot of holes in this.
For the most part, it is used to determine if the machine is using the rightsizing, etc.
The sizing recommendation will look, and say, "You are only using this at 80%," then recommend a better fit for you.
The sizing recommendations should be done in longer than two week windows. They should be looking at a whole year's worth of information because we get spikes, and once you are out of that window, it doesn't account for it. It will say that you can save money, but the reality is that it wasn't estimated with a longer running cycle and time slice. Also, if you were able to have a year long look at it, then it would be able to do some type of a linear regression model along with some predictive analytics, and say, "You spent this much this year, so we estimate with your growth rate that this is where you will need to be next year without any new features. Right now, I don't think that they do good predictions at all. Some of their competitor's do offer these predictions, so this is an area for improvement.
Therefore, we would like them to have a linear regression, so we can be predictive for budgets, allocations, and the year's follow ups. We also want to have a longer window of analytics with better certainty that our workload will fit the model, not just in a two week window.
With containers, corporate doesn't look at a container level to charge adequately, because things get masked.
One to three years.
The stability is average. Everybody knows when it is down, which is a good thing.
The scalability is average. It is not terrible, but it is not great. Sometimes things take a while to load, but any analytics that haven't ran in a while need to reboot all their stuff.
We are very big. I don't even know how many EC2 types that we have. As an example, they want us to save $12 million USD a year on budget, and I don't even know what that computes to in compute power.
I have not used technical support.
I was not involved in the integration and configuration of the product.
My team is one of the most expensive teams, and we look at it quite a bit. We have probably easily saved around $400,000 USD a year.
The product is probably not valuable until you are over a certain threshold in compute power. While I don't know what the actual cost is, if you were to say, "We could save you X amount and that would offset the cost of their product, then it is probably starting to be in the realm of being worth it."
We only use the AWS version.