AWS EC2 Spot Usage | Real-life Examples | Three Common Patterns to Handle Interruption

แชร์
ฝัง
  • เผยแพร่เมื่อ 5 ต.ค. 2024
  • EC2 Spot Instances save you a lot of money.
    We look at real-life examples of spot usage along with three common patterns for handling spot interruptions
    Scaling hotstar.com for 25 million concurrent viewers
    • Scaling Hotstar for 25...
    Topics:
    1. Why use Spot - • AWS EC2 Spot Usage | R...
    2. How to Handle Spot Interruption - • AWS EC2 Spot Usage | R...
    3. Health Check (Option 1) - • AWS EC2 Spot Usage | R...
    4. Can you lose entire capacity? - • AWS EC2 Spot Usage | R...
    5. How to handle interruption without impacting client? - • AWS EC2 Spot Usage | R...
    4. Event Subscription using Auto Scaling (Option 2) - • AWS EC2 Spot Usage | R...
    5. Persist State and Restart (Option 3) - • AWS EC2 Spot Usage | R...
    Share Your Comments, Like and Subscribe!
    Music: Stellar Wind TH-cam Audio Library
    Images: Microsoft Stock Images
    ABOUT Chandra
    Chandra Lingam teaches AWS, Security, SageMaker, ML, Regular Expression.
    Part 1: AWS Certified Solutions Architect SAA C02
    www.udemy.com/...
    Part 2: AWS Certified Solutions Architect SAA C02
    www.udemy.com/...
    Complete Course List:
    www.cloudwavet...

ความคิดเห็น • 9

  • @anishpks3354
    @anishpks3354 10 หลายเดือนก่อน +2

    It is really deep and useful you made it very interesting with console as well. Keep the good work.

  • @nagasuresh-r5m
    @nagasuresh-r5m 5 หลายเดือนก่อน +1

    Finally found some making videos on real world use cases. Nice to see your video Sir

  • @asifadam93
    @asifadam93 ปีที่แล้ว +1

    Great video 👌

  • @AzzMdA
    @AzzMdA ปีที่แล้ว +1

    Can I use spot instances as primary instances in an autodialing group?

    • @CloudConceptsByChandra
      @CloudConceptsByChandra  ปีที่แล้ว

      what is a primary instance in an autodialing group?

    • @MikePreston-darkflib
      @MikePreston-darkflib 7 หลายเดือนก่อน +1

      Generally, you want to understand your risk profile - can you deal with the (albeit) small risk of no spot instances available. For my work, we run our EKS clusters using a base load of some on-demand capacity (and specify the pods with a priority so they get scheduled in preference on available infrastructure) and then spome spot capacity to give us a large pool of cheaper resources.
      It works well, but it might take a little tuning and understanding for your application.

  • @Inficate-bd3qu
    @Inficate-bd3qu 7 หลายเดือนก่อน

    When interruption signal is sent, the spot instances have no choice but to terminate in 2 minutes right? The current jobs are not guaranteed?