Skip to content

Latest commit

 

History

History
50 lines (38 loc) · 5.38 KB

01-serverless.md

File metadata and controls

50 lines (38 loc) · 5.38 KB

Install and configure Serverless

What is Serverless?
Who could be better explaining this than Serverless:
"Just like wireless internet has wires somewhere, serverless architectures still have servers somewhere. What ‘serverless’ really means is that, as a developer you don’t have to think about those servers. You just focus on code. This was initially embodied in function-as-a-service infrastructure, like AWS Lambda, that had Serverless properties like: function as the unit of deployment, 100% auto-scaling, and pay-per-execution pricing models. More recently many other types of ‘serverless infrastructure’ have emerged such as zero configuration databases, APIs for handling user management, and highly abstracted container orchestration engines. At the end of the day, Serverless is any developer tooling which the developer does not need to worry at all about the underlying infrastructure."
You can read more about this here.

Install Serverless

During the installation process you will be asked for your AWS Access Key Id and AWS Secret Access Key you got when you created the programmatic user. Take it at hand.

  1. Follow the installation instructions only here.
  2. If you didn't clone this repo yet, do it. If you alredy done go to ./infraestructure/aws/lambda and take a look at the serverless yaml file here. There you could see the definition of your serverless service. Under the functions section you have the lambda function handler, this is the entry point for execution of your microservice. It also defines an HTTP method and path to trigger the lambda function. At the top of the file, there are some general parameters and there is also two IAM policies to allow access to parameters and encription keys needed by the lambda function to configure the database access.
  3. Take a look to the lambda function source lambda_function.py. Check or update your paremeter's name if it is necessary.

Now you are ready to deploy your serverless. Let's go!

  1. Under the directory ./infrastructure/aws/lambda/
    run:
    serverless deploy -v
    This will start to make a detailed set of instructions and will automatically create all that you need to have a working lambda function in your AWS account. At the end, if all works as expected, you will have the URL to access to your service.
    During the deployment, an API gateway endpoint will be created and linked as a trigger for your lambda function. Go to API Gateway under Networking & Content Delivery section. On the left panel you will have your API gateway which is called dev-serverless-aws-workshop. Click on it and take a look on the configuration details showed and try to find the association with your lambda. Finally, under the Stages section of your API gatway you will see dev, click on it to get the Invoke URL.
  2. Now you can test your endpoint sending a GET request to you API endpoint and passing some string to the query request parameter. As a response you will get all the articles which has that string in their title. Now is time to go to the AWS console and take a look over the results of the serverless deployment.
  3. In the AWS console, go to Lambda under Computing
  4. Go to Functions at the left.
  5. Look for you brand new lambda function (serverless-aws-workshop-dev-lambda_handler) and click on it.
  6. In the Designer section you will have a diagram with your lambda function, at the left you have the triggers (in this case API Gateway) and at the right the AWS services the function needs access to. Clicking on whichever of this you will have more details below. Take some time exploring this.
  7. Finally you could test your lambda function from the console with the button test at the top right.

Maybe you haven't access to your RDS because previously in the workshop you created a VPC where you put the DB. Your lambda function are outside the VPC and your DB haven't public access (from outside of your VPC). One solution could be having your lambda inside your VPC but this has some drawbacks related to the long warming time required by the lambda function to be ready to run on each call. For this workshop we will put our RDS in the default VPC and subnets accessible from outside and then will give public access to de DB. Be careful and don't do that in a production environment! There is other ways to workaround this.

Give public access to your RDS.

  1. Go to RDS under Database and then to Databases on the left.
  2. Select your database and click on Modify.
  3. Under Network and Settings select default Subnet group.
  4. Select a security group with allowed TCP traffic from 0.0.0.0/0 to TCP Port 5432. If you havent this security group created you can create one in VPC security groups under EC2 section addint a Inbound rule with type PostgresSQL.
  5. Then select Yes on Public accessibility.
  6. Click on Continue
  7. Select Apply Immediately and click Modify DB Instance

At this point you could make GET request to your new serverless API endpoint using the URL associated during the deployment and configured in AWS API Gateway.

Next: Connect your load balancer to your lambda.