EC2 - Load Balancing and Auto scaling

Load balancing and Auto scaling


Let's suppose we have deployed our web application to the EC2 instance. And all of the sudden, our app. got viral and previously 10k daily users became 1 million. So we want our infrastructure to use less resources when there is little or no user, more when there is high load or spike, it somehow must adjust itself in a way to withstand to this load. This is when auto scaling comes into play.



Load balancer - is a proxy that allows you to distribute connections to servers. It provides fault tolerance and high availability. Load balancer can be front of  EC2,ECS, other load balancers, lambda functions etc.

There are 3 types of load balancer in AWS:

  • Application load balancer (operates at Application layer (L7) http, https etc.)
  • Network load balancer (operates at IP layer (L4) -  TCP, TLS, UDP etc.) 
  • Gateway load balancer (uses Geneva protocol, balances firewalls,IDS/IPS, operates at layer 3 )


Auto scaling - allows you to automatically and terminate EC2, ECS, EKS instances based on your needs. It maintains availability allows you to scale easily. We need to define our EC2 template, so each time when auto scaler needs to create new EC2 instance, it will use that. And there is one more thing I want to add on top of that, auto scaler needs some information when to create, when to terminate instances. Based on the CloudWatch data and auto scaling policy we define, it will scale the system.

Types of Auto scaling:

Manual - make changes to AutoScaler manually
Dynamic - automatically scales based on demand (request size, CPU usage etc.)
Predictive - uses Machine learning to predict 
Scheduled - scales based on schedule (date/time)



Let's create auto scaler and connect it to the elastic load balancer

  1. First let's create own security group which will be very handy later on


  2. Add allow 80 and 22 ports to inbound rules.





  3. After that we should create a launch template which auto scaler will use it to create EC2 instances.








  4. Use amazon AMI and don't forget to select our security group we just created. 











  5. Add the script we used to create simple webpage in previous tutorial (https://www.blogger.com/u/1/blog/post/edit/6922501496740215722/8300358555009620696)






  6. Let's create auto scaling group.






  7. Use the launch template we just created.




  8. Adjust the minimum and maximum capacity based on the need




  9. We need to create a target group, so that load balancer will know which targets it will redirect traffic to. And for the auto scaler where to register targets to.




  10. Specify target group details



  11. And lastly we should create load balancer itself.




    Here select the name of target group !


  12. Okay, now let's connect our load balancer to auto scaling group too.






  13. Finally we are able to see results of our efforts here, 
























Comments

Post a Comment

Popular posts from this blog

Identity Access Management - How to create user in IAM ?!

Image
Introduction IAM stands for Identity and Access Management. IAM is essential component of AWS allows you to create and manage access of user, role, application and control authentication/authorization. There are 4 essential components of IAM :  Users - are individuals or applications want to access to your AWS resources. Groups - are collection of users. By this you can assign permissions to group instead individual users. Roles - are special access type in IAM that allows user/resource to access another resource via temporary credentials. For example EC2 instance can access to DynamoDB or S3 bucket resources without using permanent credentials. This is possible only via Role which is great security point of view. Policies - are JSON documents that define permissions. In policies we define what actions are allowed or should be denied. This can be attached to users, groups, roles. Hands on IAM exercises :  Login to your AWS account and search for IAM :  After clicking IAM icon, you wil
Read more

AWS pricing fundamentals

Image
When using Amazon Web Services (AWS), there are three primary factors that determine your costs: Compute : This includes the amount of resources such as CPU and RAM you use, as well as the duration for which you use them. Data Storage : The quantity of data stored or allocated. Outbound Data Transfer : The amount of data transferred out from all services. It's important to be cautious and monitor your usage to avoid unexpected costs. AWS Pricing Models On Demand (Pay as you Go) The On Demand model allows you to: Easily adapt to changing business needs. Adjust based on your requirements. Reduce the risk of paying for unused capacity. Reserved Instance (Save When You Reserve) By investing in reserved capacity, such as with EC2 or RDS, you can: Ensure you have resources available for unexpected scenarios like web traffic spikes. Reserve resources for 1-3 years, which can lead to significant discounts. Save up to 75% compared to the Pay-as-you-Go model. The more you pay upfront, the gr
Read more

IAM - Roles

Image
IAM Roles Roles in AWS has special meaning. With this functionality you can define different variety of roles and assign it users or your application/service can use them. Roles are specifically important for AWS security. Because roles use temporary credentials with services. So that we don't need to use own permanent credentials to do some operations. If somehow hackers got access to AWS infrastructure they won't able to access our cached permanent  credentials. Here are some of advantages of using roles : Reduced Risk of Credential Exposure Least Privilege Principle Cross-Account Access  Let's create a role and assign it to the user  Click to "IAM" , "Roles" and then "Create role" button.  We need ARN (we talked about this in the previous post) of the user. Copy 12 digit unique ID. During role creation select "AWS account" since we are planning to add this role to user. Enter the ID we got from previous step. This ID defines our us
Read more