This is simple but excellent. AWS should have graphs like that on their calculator page to help people decide how long it will take for their investment in a reserved instance to get ahead of on-demand (or another reserved instance plan).
Perhaps it's pessimistic of me to say so, but I doubt Amazon wants you to be able to appreciate the price relationship with quite this degree of immediacy. Sure, the data is there for you to plot yourself or use a third-party tool such as the OP's. But if you naively select an on-demand instance (the default) where a dedicated instance would be more advantageous for you, Mr. Bezos is made a tiny bit happier.
Amazon actually regularly sends you e-mails letting you know if your configuration can be optimized to reduce it's price by moving to reserved instances. This suggests to me that they favor the reduced revenue but predictable capacity planning of reduced instances.
Maybe, but the heavy reserved instances come with a required hourly commitment. If you buy one, you have to pay the 24/7 hourly usage fee regardless of whether you're actually running a server. That strikes me as the bigger "gotcha."
There is one more "gotcha" if you dig deeper. That is the drop in EC2 prices and Amazon in general has been vague about passing on the benefits of price drop to old reservations ( https://forums.aws.amazon.com/message.jspa?messageID=325153 ) . We even thought about adding speculative price drop and then do the comparison, but dropped the idea. It will be interesting to hear the take of HN community on this ..
I'm not sure how these calculations are correct - I have eu-west-1 set up, and the price for 1yr heavy reserved is fixed. But the AWS pricing still has an hourly rate against it. The final cost looks right, but the initial cost (eg., taking out the reservation but never running the instance) should just be the setup cost I think?
When you reserve an instance, it doesn't entirely eliminate the hourly cost to run it. What you are essentially doing is paying a large chunk up-front in exchange for a heavily discounted hourly rate, which ends up saving you money by the end of your 1 - 3 year contract.
Quote from the page above
"
Light and Medium Utilization Reserved Instances also are billed by the instance-hour for the time that instances are in a running state; if you do not run the instance in an hour, there is zero usage charge. Partial instance-hours consumed are billed as full hours. Heavy Utilization Reserved Instances are billed for every hour during the entire Reserved Instance term (which means you’re charged the hourly fee regardless of whether any usage has occurred during an hour). "
maybe he/she was assuming if you are going for heavy reservation you will be using 100% of hours? the other reserved instance pricing changes with hours
havent done a check to see if he was taking into account the discounted hourly price or just adding on on-demand hours...
Unfortunately, it's not an assumption. It's not immediately clear when looking at the pricing, but if you buy a heavy reserved instance you have to pay the 24/7/365 hourly rate. It's something for startups to consider as they look at their burn rate.
I like the idea, it's just a bit difficult to use for what I'm getting out of it. Simple things like making clear what the "number of days" slider means (it didn't take long to figure it out, but I had to use the context to draw my conclusion). I'm used to seeing my instances in the typical "t1.micro, m1.small, m1.medium, etc" format, so I was a bit thrown off by the "Instance Category" dropdown.
It would also be nice to have values on the x axis of the graph so I can see exactly when certain thresholds are crossed - as of right now I'm just assuming that the max x-value is my specified days * 24.
Again, cool idea, but I'd be more inclined to hit up the AWS Cost Calculator (http://calculator.s3.amazonaws.com/calc5.html). It's ugly, but it allows me to quickly add many resources from all of the AWS services for a total monthly cost.
Thanks for the feedback. We are doing the "t1.micro, m1.small .."translation right now. (Edit : changes pushed)
X axis value seems to be some browser issue. On some browsers it shows up only at reduced zooms. Will fix it.
Also, we are not trying to replace AWS cost calculator. It was more like putting our own dog food out to answer the question, that at what point/usage is the reserved instance better than on demand.
Cool tool. One suggestion, in the dropdown where you list the types of instances, can you append the abbreviations that AWS uses (e.g. M1-small, M2-medium, M3-large). I'm sure I'm not alone in referring to instances by those codes rather than their full names, which I seem to always forget.
If you're taking feature requests, it'd be really handy to see pricing breakdowns for Windows instances as well. They have a different breakeven point than the Linux ones.
I second the request for Windows instances. I currently use a spreadsheet to track everything, then update the costs every few months--comparing instance sizes with OS's. It seems like something Amazon should offer themselves, but I'm glad you're taking on the challenge. :)
That very cool, something that many of us will find useful.
Feedback: The top bar folds over when the browser window is resized (looks like anything less than 1024 px). The page then starts to overlap on itself blocking out content at the top.
For 30 days 24hr scenario, "Total cost of instance" graph is showing bogus numbers on the iPad. Let me know if you are unable to replicate and need a screenshot.