<p><img src="https://s3.us-east-1.amazonaws.com/resource.mixa.site/meric/posts/img/indexing-dynamodb-items-to-elasticsearch-using-aws-lambda/cover.jpeg" priority="1"></p>
<p>DynamoDB is Amazon’s NoSQL database which offers single-digit milliseconds latency. It is great for variety of use cases, but when you need to run complex search queries on your dataset, you quickly realise it is not designed for it.</p>
<p>You can try using composite primary keys, local and global secondary indexes to fulfil your needs. But as the queries get more complex, you might realise those are not enough. But don’t worry, you are not alone. Lots of people who have faced with this problem chose a similar solution; using ElasticSearch for their complex search operations while keeping the DynamoDB as the authority for the data.</p>
<p>Here we are going to learn how ElasticSearch can be plugged into your DynamoDB with a click of a button using CloudFormation.</p>
<p>The same steps can be done using AWS Web Console, but I think making use of CloudFormation is better as often times you need to create the same resources for multiple stages and regions.</p>
<h2 id="creating-dynamodb-table"><a aria-hidden="true" tabindex="-1" href="#creating-dynamodb-table"><span class="icon icon-link"></span></a>Creating DynamoDB Table</h2>
<p>Let’s start with creating a DynamoDB table using CloudFormation. Here we create a table called ‘OrderTable’ whose key is orderId. It also sets read/write capacity to 5. One of the most important things here is that we also enable DynamoDB streams. Whenever an entry is created or updated, it will be streamed automatically. It will only return the new image, but you have <a href="https://docs.aws.amazon.com/amazondynamodb/latest/APIReference/API_StreamSpecification.html">other options</a> if you also need the old image or just the keys.</p>
<pre><code>OrderTable:
  Type: 'AWS::DynamoDB::Table'
  Properties:
    AttributeDefinitions:
      - AttributeName: 'orderId'
        AttributeType: 'S'
    KeySchema:
      - AttributeName: 'orderId'
        KeyType: 'HASH'
    ProvisionedThroughput:
      ReadCapacityUnits: 5
      WriteCapacityUnits: 5
    TableName: 'OrderTable'
    StreamSpecification:
      StreamViewType: 'NEW_IMAGE'
</code></pre>
<p>If you already have an existing table and don’t want to create a new one, you can get its SourceArn from AWS Web Console, and use it in the following sections.</p>
<h2 id="adding-a-lambda-function"><a aria-hidden="true" tabindex="-1" href="#adding-a-lambda-function"><span class="icon icon-link"></span></a>Adding a Lambda function</h2>
<p>Here we print the records coming from DynamoDB stream. We also add a Role to access various AWS resources from our lambda function which we’ll cover later. We use ZipFile which allows adding inline code to our CloudFormation template.</p>
<pre><code>EsIndexerFunction:
  Type: 'AWS::Lambda::Function'
  Properties:
    Handler: 'index.handler'
    Runtime: nodejs6.10
    Role: !GetAtt LambdaRole.Arn
    Code:
      ZipFile: |
        exports.handler = (event, context, callback) => {
          event.Records.forEach(record => {
            console.log(record);
          });
        }
</code></pre>
<h2 id="mapping-between-the-stream-and-lambda-function"><a aria-hidden="true" tabindex="-1" href="#mapping-between-the-stream-and-lambda-function"><span class="icon icon-link"></span></a>Mapping between the stream and lambda function</h2>
<p>This defines the mapping between DynamoDB Stream and our Lambda function. Whenever there is something new in DynamoDB Streams, it will trigger our lambda function with those records.</p>
<pre><code>TableStreamLambdaMapping:
  Type: 'AWS::Lambda::EventSourceMapping'
  Properties:
    BatchSize: 2
    EventSourceArn: !GetAtt OrderTable.StreamArn
    FunctionName: !GetAtt EsIndexerFunction.Arn
    StartingPosition: 'LATEST'
</code></pre>
<h2 id="lambda-iam-role"><a aria-hidden="true" tabindex="-1" href="#lambda-iam-role"><span class="icon icon-link"></span></a>Lambda IAM role</h2>
<p>Here we give our lambda function access for writing logs so that we can check them in CloudWatch and also access for reading DynamoDB streams.</p>
<pre><code>LambdaRole:
  Type: 'AWS::IAM::Role'
  Properties:
    AssumeRolePolicyDocument:
      Version: '2012-10-17'
      Statement:
      - Effect: 'Allow'
        Principal:
          Service: 'lambda.amazonaws.com'
        Action: 'sts:AssumeRole'
    Path: '/'
    Policies:
      - PolicyName: 'LambdaRolePolicy'
        PolicyDocument:
          Version: '2012-10-17'
          Statement:
          - Effect: 'Allow'
            Action:
            - logs:CreateLogGroup
            - logs:CreateLogStream
            - logs:PutLogEvents
            Resource: 'arn:aws:logs:*:*:*'
          - Effect: 'Allow'
            Action:
            - dynamodb:DescribeStream
            - dynamodb:GetRecords
            - dynamodb:GetShardIterator
            - dynamodb:ListStreams
            Resource: !GetAtt OrderTable.StreamArn
</code></pre>
<h2 id="adding-elasticsearch"><a aria-hidden="true" tabindex="-1" href="#adding-elasticsearch"><span class="icon icon-link"></span></a>Adding ElasticSearch</h2>
<p>This creates a t2.micro ElasticSearch instance which is included in Free Tier. It also allows LambdaRole to access to the instance.</p>
<pre><code>ElasticsearchDomain: 
  Type: 'AWS::Elasticsearch::Domain'
  Properties:
    DomainName: 'es-order'
    ElasticsearchClusterConfig: 
      InstanceType: 't2.micro.elasticsearch'
      InstanceCount: 1
    EBSOptions: 
      EBSEnabled: true
      Iops: 0
      VolumeSize: 10
      VolumeType: 'standard'
    AccessPolicies: 
      Version: '2012-10-17'
      Statement: 
        - Effect: 'Allow'
          Principal: 
            AWS: !GetAtt LambdaRole.Arn
          Action: 'es:*'
          Resource: '*'
    AdvancedOptions: 
      rest.action.multi.allow_explicit_index: 'true'
</code></pre>
<h2 id="updating-lambda-to-index-documents"><a aria-hidden="true" tabindex="-1" href="#updating-lambda-to-index-documents"><span class="icon icon-link"></span></a>Updating Lambda to index documents</h2>
<p>Now let’s update our previous lambda function to index DynamoDB stream records to ElasticSearch. Note that we need to sign our requests, otherwise you’ll get an authorization error.</p>
<p>Consider using <a href="https://github.com/TheDeveloper/http-aws-es">http-aws-es</a> library which makes things quite easy if you are using S3Bucket and S3Key instead of ZipFile. It allows you to use <a href="https://github.com/elastic/elasticsearch-js">elasticsearch-js</a> client and also handles the request signing part.</p>
<pre><code>EsIndexerFunction:
    Type: 'AWS::Lambda::Function'
    Properties:
      Handler: 'index.handler'
      Runtime: nodejs6.10
      Role: !GetAtt LambdaRole.Arn
      Environment:
        Variables:
          ES_ENDPOINT: !GetAtt ElasticsearchDomain.DomainEndpoint
          ES_REGION: !Ref AWS::Region
      Code:
        ZipFile: |
          var AWS = require('aws-sdk');
          var path = require('path');
          var creds = new AWS.EnvironmentCredentials('AWS');

          var esDomain = {
              endpoint: process.env.ES_ENDPOINT,
              region: process.env.ES_REGION,
              index: 'test',
              doctype: 'order'
          };
          var endpoint =  new AWS.Endpoint(esDomain.endpoint);

          exports.handler = (event, context, callback) => {
            event.Records.forEach(record => {
              postDocumentToES(record.dynamodb.NewImage, context);
            });
          }

          function postDocumentToES(doc, context) {
              var req = new AWS.HttpRequest(endpoint);

              req.method = 'POST';
              req.path = path.join('/', esDomain.index, esDomain.doctype);
              req.region = esDomain.region;
              req.body = JSON.stringify(doc);
              req.headers['presigned-expires'] = false;
              req.headers['Host'] = endpoint.host;

              // Sign the request (Sigv4)
              var signer = new AWS.Signers.V4(req, 'es');
              signer.addAuthorization(creds, new Date());

              // Post document to ES
              var send = new AWS.NodeHttpClient();
              send.handleRequest(req, null, function(httpResp) {
                  var body = '';
                  httpResp.on('data', chunk => body += chunk);
                  httpResp.on('end', chunk => context.succeed());
              }, function(err) {
                  console.log('Error: ' + err);
                  context.fail();
              });
          }
</code></pre>
<p>So at this point we have everything we need. All the new DynamoDB data is being indexed to ElasticSearch behind the scenes.</p>
<p>You can also see the full CloudFormation template <a href="https://gist.github.com/merictaze/44cb99335300fb1121512eb9beea3ab3">here</a></p>
<p>Feel free to comment below if you have any questions or feedback.</p>

Maybe you’re just getting started with distributed systems and system design, or you just need a quick recap for your upcoming interview. In either case, in this post, you’ll find the most common concepts of system design and the most important aspects of them. Let’s get started.

Recap of System Design Interview Concepts

Although what react is doing looks quite complex, underlying logic is very simple. And the easiest way to understand is just running some simple examples. We’ll dive into how react renders components and updates the actual DOM.

How React Rendering Works - A Step by Step Guide

DynamoDB is Amazon’s NoSQL database which offers single-digit milliseconds latency. It is great for variety of use cases, but when you need to run complex search queries on your dataset, you quickly realise it is not designed for it.

Indexing DynamoDB Items to ElasticSearch using AWS Lambda

Multitasking is the source of all evil. Until recently, I also liked having multiple screens and multi-tasking like many others. It was great to code on one screen while reading the slack messages and emails on the other screen. But, it was so easy to get distracted by continuous incoming notifications while trying to read, write or code something. Often, I was feeling tired in the middle of the day because continuous context switching was draining my energy and destroying my focus on the task I was working on. 

5 Tips To Stop Multitasking and Double Your Productivity

In this post, we are going to setup an environment with API Gateway, Lambda and DynamoDB using serverless framework which can be deployed to AWS directly or run fully local.

Going Serverless Offline

You solved 60 LeetCode hard and 300 LeetCode medium questions, you were a rockstar in your coding round and provided the most efficient solution in the world, but you couldn’t get the offer. Sounds familiar? Either case, keep reading, you might find some useful tips.

How to Succeed in Your Coding Interview

While making a simple request to https://google.com to see that simple search page, there are a lot of things going on behind the scenes to keep you secure. Understanding this flow will give you a lot of confidence in various topics like Digital Certificates and Signatures, Symmetric/Asymmetric key encryptions, Certificate Authorities and how they all work together to help you have secure communication.

What happens when you hit google.com using HTTPS

We are going to create a single component that supports text, textarea, select, radio, and checkbox. It will absorb all implementation differences for different inputs and will allow us using a unified component in our Formik forms.

Creating a unified Formik input field to support all input types seamlessly

System design questions can be asked for any level with different expectations from the candidate, but to get a Senior offer, you have to rock this round. Practice makes perfect - like for anything else, it applies for system design interview too. So, keep practicing without taking shortcuts, practice as if it’s the real interview and always follow the same steps to make it perfect.

How to Succeed in Your System Design Interview

Attacks like CSRF or XSS are still not clear to many of us. It is not because they are super hard to understand, but it requires some basic understanding of concepts like SOP, CORS, CSP, HTTPS. Let's start with what's Same-origin Policy