Recently, I wrote about creating AWS EC2/Compute On-Demand Capacity Reservations creation and how to launch & use instances of that capacity reservations. Please give a read here Link
Welcome to a post where we can see When & How to gain on cost with AWS On-Demand Capacity Reservations !!
Ideally AWS Capacity Reservations' main purpose of existence is to reserve the 'Instance Type' & 'Capacity', as needed
Other Reasons Include
1) Business needs a certain EC2 Instance Type but needs flexibility
2) Use case cannot commit for usage of instance and hence not going for Reserved Instances feature
3) Organization wants the Architecture to have high resiliency and cannot risk on the availability of particular instance type
4) Requirement is adhoc like any season or event demands and this is likely to come down with a short duration of time
5) ODCR - can be cancelled any time
Risks in using ODCR:
1) There is no assurance that you will get instance
2) Billing will start based on the On-demand rate of the EC2 instance, though it is being run or active or EC2 exist
Below is the note on ODCR capacity reservation creation.
Let us take an example with below ODCR
1) Above is sample of ODCR
2) Now I will be billed with On-Demand rate on enablement of ODCR request
3) In order to save money on this ODCR situation, I can create a AWS Savings Plan of specific instance family under 'EC2 Savings Plan' and give an hourly commitment and use the same date of start date as in ODCR request
Sample AWS Savings Plan Request
The only limitation is setting a commitment of 1 year in purchasing a savings plan but this can be planned if the usage can be matched with some other application requirement, if it could fit
Think through and apply the strategies to stay optimized !!