Need advise regarding scope for cloud cost monitoring & optimization tool.
by•
I have been thinking about developing a tool that can show statistics like below to show for all my cloud or any infra that are hosted for a company. I know each cloud have their own way to show this but do we already have a common tool for this, that can show the details for all servers(multi-cloud)? Will this be a worthy product? Can someone give me some suggestions please.
"Here's what changed this month."
+2 services ADDED : analytics-01 & app-02 came online.
WASTE: app-01 is stopped, still billing — $374/mo.
RIGHT-SIZE: db-01 idles at 16% CPU p95 — downsize.
76 views
Replies
I love the “what changed this month” idea. i'd much rather see the few things I need to fix than dig through another billing dashboard.
@tessa_lynch A big plus for me would be alerts that explain why a cost changed instead of just showing another dashboard full of numbers.
@tessa_lynch @isabella_wandrei Hello, i like the focus on actionable suggestions here. If the tool can connect the cost increase directly to a service or infrastructure change, I could see teams checking it regularly.
@ludovica_eleazer That is the idea here. If you look at my post, I tried to give as much as info possible to take action on the items mentioned. The tool can connect to multi-cloud or any server hosting and get this data. This is vendor agnostic tool. Thanks for the feedback.
@isabella_wandrei Thanks for the feedback. My current vision is to generate PDF report and send it to configured mail ids every month. Can add a notification/alerts via specified communication modes.
@tessa_lynch Thanks for the feedback. Please let me know if you need anything else to be shown along with these details.
One suggestion from me would be to add a confidence level to recommendations. I’d want to know whether a resource is definitely oversized or if the system is making a best guess from limited usage data.
@elsie_ramsey Thanks for the feedback. Sure we can add confidence score/tag against each suggestions based on the range of data. I think the longer the server usage data, the higher the confidence. Hope this kind of logic will work.