Understand Opensearch Reserved Instances and Savings Plan

0

Hi, I am trying to understand 1) Are EC2/Compute saving plans applicable to the instances in use by an OpenSearch domain 2) Are the reserved instances purchased through for OpenSearch convertible or standard instances? 3) What is better way to reduce the cost - reserved instances or savings plan - for an OpenSearch domain that has an ever increasing traffic that might require us to scale up the instances (for example we reserved initially 3 data nodes of m5.large.search and later want to replace these nodes with m5.xlarge.search). Any insights are highly appreciated.

1 Answer
3
Accepted Answer
  1. Are EC2/Compute saving plans applicable to the instances in use by an OpenSearch domain

Edit: No, compute savings plan won't be applied for opensearch. Refer the comment section.

  1. Are the reserved instances purchased through for OpenSearch convertible or standard instances?

Please refer answer to question 3. RIs for opensearch are not flexible.

  1. What is better way to reduce the cost - reserved instances or savings plan

It's a subjective question, in theory, RI provide better discount but if you are not sure about instance family and may anticipate that to change frequently, then RIs would not be better fit.

In case of Opensearch: RI are the only option as CSP doesn't cover opensearch. Amazon OpenSearch RIs are not flexible; they only apply to the exact instance type that you reserve. For example, a reservation for eight c5.2xlarge.search instances does not apply to sixteen c5.xlarge.search instances or four c5.4xlarge.search instances. Take a look at Reserved Instances in Amazon OpenSearch Service

Enter image description here

References for opensearch:

General reference:

Hope you find this helpful.

Comment here if you have additional questions, happy to help.

Abhishek

profile pictureAWS
EXPERT
answered 9 months ago
profile picture
EXPERT
reviewed 2 months ago
  • Thanks for the detailed answer. One particular issue that is bothering me about reserved instances for OpenSearch is that they need to be purchased on OpenSearch console (these are not accessible from the EC2 console) and on OpenSearch console there is no option to choose Standard vs Convertible. That is the source of the confusion whether these reserved instances are standard or convertible.

  • I'm sorry, I didn't specify that. I should rather mention this in the answer itself, however I had included the reference "Are OpenSearch Reserved Instances automatically applied to running on-demand OpenSearch instances?", which talks about this and also attached the respective pricing document. Amazon OpenSearch RIs are not flexible; they only apply to the exact instance type that you reserve. For example, a reservation for eight c5.2xlarge.search instances does not apply to sixteen c5.xlarge.search instances or four c5.4xlarge.search instances. https://docs.aws.amazon.com/opensearch-service/latest/developerguide/ri.html

  • I've updated the answer for details specific to opensearch. Hope you find that helpful. Comment here if you have additional questions, happy to help.

  • I'd like to make a correction. Compute Savings Plans DO NOT apply to OpenSearch. Non-EC2 services DON'T have Savings Plans, and can ONLY benefit from Reserved Instances. The RIs for non-EC2 services are service-specific, and indeed have no concept of Standard vs Convertible. Think of all RIs for non-EC2 services as Standard. OpenSearch RIs are also not size-flexible, and can apply to resources only if they match the attributes exactly (region, resources type, generation, family, and size). So, there can't even be consideration whether you should choose between SP or RI in this case.

  • Thank you Nataliya for pointing that out. I just corrected that. @testaws123: As Nataliya mentioned, CSP won't be applied for opensearch and only RI is the option.

You are not logged in. Log in to post an answer.

A good answer clearly answers the question and provides constructive feedback and encourages professional growth in the question asker.

Guidelines for Answering Questions