Hello everyone, welcome to my presentation on unleashing the power of serverless for building scalable and cost solution using AWS Lambda and GQGen. In this presentation we are going to cover two most powerful technologies. One is serverless computing and another is one of the go popular library GQ gen for generating graphqL apps. We'll start by exploring what is serverless computing. Then we'll go in more deep dive about how we can implement using AWS lambda and then we are going to cover what is graphQL APIs and how we can implement using gqengine. Before going further, I just want to give a small introduction about myself. My name is Durgaprasad Budhwani. I'm working as a chief technologist at Tech nine throughout my career. I always fortunate enough to work with latest technologies, with latest trends and with various clients with multiple domains and I always fortunate enough to provide them a cutting edge solutions which fulfill their needs. And this is what we do in tech nine. We always work on our latest technologies. Let's talk about serverless computing in case of serverless computing which is also known as function as a service, the cloud providers such as AWS, Azure, Google or Versaille will manage the infrastructure and automatically allocate resources for running and executing the code in the response of any request or can event. In case of traditional computing, the developer has to manage the servers. It could be a developers or a DevOps team has to manage the servers. With serverless computing, the developer only need to write the code for the specific function they want to run and the cloud provider will handle the infrastructure, scalability and availability of those functions. This allows the developer to focus on writing the code and building the application without worrying about the underlying infrastructure. There are lot of advantages of serverless architecture or serverless computing. One of the main primary advantage of serverless computing is a cost reduction. Since users only pay for the actual amount of compute time consumed by their functions, they can save money by avoiding the cost of managing and maintaining servers. Serverless computing can automatically scale to handle changes in demand. As the number of requests or event increases, the cloud provider can allocate more resources to handle those workload, ensuring that the application remains responsive. Serverless computing allows developer to focus on writing code and building application rather than managing infrastructure. This simplifies the development process as well as developer can only avoid the complexity of managing servers, load balancers or other infrastructure. Component serverless computing allows developer to focus on writing code and building application rather than managing infrastructure which include it actually simplified the development process, the developer has not to worry about managing servers, load balancers, other infrastructure component with the serverless computing, developer can quickly and easily develop code changes without worrying about the underlying infrastructure. This can speed up the development process and improve time to market for new features and application and serverless computing can improve the application reliability by automatically handling tasks such as server maintenance, load balancing and scaling. This can help to reduce the risk of application downtime and ensure that the applications remain available to the users. AWS Lambda is nothing but it's a serverless compute service provided by Amazon Web Services that allows developer to run the code in the response of event or request. With AWS Lambda, the developer can write code in variety of programming language which include Java, Python, node, JS, and then they can upload their code to run as a function. In AWS environment, these services automatically handle the compute resource scaling, availability of functions so that the developer can focus on writing code and building application rather than managing servers. AWS Lambda support variety of programming languages there are a couple of languages which has provided as a native which include Java net. If a programming language is not supporting AWS lambda, but you can also create a docker image or you can also create your custom runtime and then you can also run your code on AWS lambda which is basically happened when you are choosing a PHP language. You can create your docker image or you can use a custom runtime. In case of the trust language. AWS lambda is often used for the event driven application where code is triggered in the response of events such as changes in Amazon S three or any new message which is pushed into AWS SQS, Amazon Q service or in case of any message which has been pushed to Amazon simple notification. It can also be used to create API and run bad jobs and perform data processing tasks. So we are going to see how we are going to write a simple go application for AWS lambda. But before going to more on how we are Golang to write this for the AWS lambda, first we'll take a simple example of go application which is listening to any HTTP request. For that we can see a simple example where a main is kind of starting point. This program is using gin framework. Jin is nothing but one of the popular web framework. Here it is creating a router and the router is then listening to an HTTP request, a get request. When someone is going to put for this specific request it is going to return a hello world and then at the end this web framework will start listening to HTTP request on the port number 80 80. That means when someone is going to hit the localhost HTTP localhost 8080 request, this will going to print the hello world. A very basic example in Golang go application. Now when you want to write a similar application for the AWS lambda, in that case, now when we want to port the similar application for this AWS lambda, then we need to do a couple of more changes. The same thing is we are going to use for the gin framework for the routing. We need to import multiple libraries like AWS lambda go events, AWS go lambdas. And then we are going to also going to use this lambda go API proxy. So this proxy support for multiple framework, it support for fiber, it's support for gorilla, it's one of the library provided by the AWS lab so that it can use the existing router, the router which has been created, and it can kind of mostly port or wrap as per the AWS lambda. So we're doing the same thing at the global level. We are just initializing this gin lambda. We are doing the initialization of part this thing. So one of the, we can say a small drawback of AWS lambda is because it's going to run a function periodically. So function initialization time is going to take some time. So that is called as a call start time. We can kind of reduce the overall call starter for the successive calls. And that's why so the calls which required to do one time initialization, it's recommended to put into the initialization block so that for the next successive call this block will not kind of reinitialize. So we are just doing the same thing here in the initialization bar. We are just initializing the router. We are also kind of saying that if someone is going to hit for this forward slash endpoint, just return hello serverless. And then we are initializing our gin lambda which is a proxy and we are passing the router. So this handler will be kind of starting point for the AWS lambda. It accepts few things. One is context, which has a lot of information. It has information about the x ray. If you're enabling the x ray, it has information about from where the call has been initiated, kind of a lot of information in the context. And then it just provide another thing which is a proxy API request proxy. Then this proxy will have information about the exit path that it is calling. Then we are just kind of wrapping inside the gin lambda proxy with context where we're passing both the things and then the starting point will be lambda start. So it's like a same like it will wait for the request, then it will call the lambda start, it will call the lambda. We already have a router which is already initialized. We are good to go. And when someone is going to hit this URL, it's going to call hello serverless. So we have seen that lot of code, for example, for the router AWS, well as when you are writing a basic goal and application, most of the code is same. Only a small kind of a wrapper is required when we are running this code for the AWS lambda. So let me quickly open my editor and just walk you through on how we can kind of make sure that we are trying to use the same code. And I will also walk you through what are the practices which AWS recommend when we are going for the AWS lambda. Now I'm going to show you that how we can use the code sharing. That means the code which has been written for serving the AWS lambda. We can also run the same code locally. One of the main challenge of running any of the service technology code local is how can test everything locally. So there are a lot of tools available. One of the great tool especially for AWS is called as local stack. By using that local stack you can also run AWS lambda locally. Internally it just deploy into this local stack environment which is kind of equivalent to AWS lambda environment. And by that mechanism we can test it out. But again, kind of deploying that code, it's kind of a time consuming process. So this is a simplified process which I am figured out and I hope that you are also going to get some benefit of that thing. So the solution is pretty simple. We need to identify what is the common code. This is also one of the combination from AWS lambda is to keep your initialization code separate from your business logic code. So in our case, consider that our business logic is to serve our HTTP request. In that case, we are just kind of moving that code, especially for the routing code, into a separate file. Let's call as router go and then on our local go which is going to run on the local system, we're just mentioning here that just call the router and just initialize the router. It was the same code which I shown a few seconds back in the presentation that we can run a basic lambda code, just initialize the gen router and run the code. So this is good. We can quickly test it out whether this is working or not. I'm going to run the same code. Go run local code which is pointing this local folder and it has started listening on the port number 80 80. I was just using one of the feature of I'm using this Golang tool for all this kind of a development purpose, and this is one of the good tool, especially for the Go language. You can also use the same tool for the Vs code, the same principle which I'm showing here. It will be work and I also request the organizer to share the GitHub link along with the presentation. So now the port is running on 8080. You can see I can directly call a request which is our get HTTP request, and we can get the data which is hello world which is coming from this router. Router has been initialized and here we can get the hello world thing. Similarly, if I want to write the code for lambda and I need to kind of follow the same principle, that the initialization code will be separate and the business logic code will be different, then this is how it's going to be happened over here. Again, the main is kind of a starting point for us. Then we are initializing everything. We are taking the router information in our goal and initialized. We are just initializing the router. Then we are also using AWS lambda proxy. Inside the proxy we actually created a global variable. Here we are just passing, just creating the kind of a global object new adapter and passing the router. And then this is the starting point of AWS lambda, especially for the HTTP request. We got the context, we get the request and this will be the response. And since we are using this AWS labs gin proxy, it just kind of wrap everything and the same code will work for both the things. Now this is how we are able to kind of share the code between our local thing and lambda, and we can kind of run this entire code too. There are various options for the deployment, but for this presentation I'm going to use serverless application model AWS SAM for the deployment. So the other options could be we can deploy using a terraform, we can deploy using cloudformation, we can deploy using plume, or we can deploy using AWS Cdk. But if you are going to do for some sort of PoC, then this SAM will be kind of much easier. For the deployment. We need to install a SAm ClI which is equivalent to our AWS ClI. After installing same CLI we need to run a command which is same initialize. So this is going to ask you a couple of things. It has a couple of boilerplate project for the Golang, for rust language, for node js, multiple languages and we can also choose a custom project also. So I already installed this SAM initialize and it has created kind of some project. So after running a SAM initialization part, it has created some sample project for me. I actually taken the respect information which is required to run this application does include a Mac file where we are going to run a SAm build command template yaml file which is responsible for kind of. It is mostly equivalent to a cloudformation template, but it's mostly configured based on Samway so that we can deploy our lambda. And a couple of modification has been made so far. And this is kind of one of the important thing for us is because this will tell us what kind of thing we need to do. So it start with a folder location. So when I'm going to run my build command, which is my back build command, it's going to generate the artifact. But this build command will utilize this same template. It's going to see where the code path is. In our case the lambda server go, that was where the code was. There it will create the name of the handler. So this handler will be kind of a logical name for the handler then which runtime we need to use. So the lambda support multiple runtime node js Python. So similarly right now it's supporting go one x version, then you can select an architecture. So AWS Lambda support two kind of architecture. One is we get called AWS 86 64 and another one is arm. For this I'm just selecting can arm. Then it will ask for when we are going to invoke this lambda, when this lambda is going to be called. So here we are saying that this event will have a logical name, catch all. It's just kind of a logical name. And then the type of event will be an API, which is kind of can API request from your API gateway, and it's going to create an API gateway for us. And when someone is going to say a specific endpoint get request, then this is going to invoke at the output of this one. When I'm going to run Sam build command and send deploy command is going to generate the entire structure stack for me. And once the structure is ready, then we are going to see a couple of URLs at the end. So what I'm saying to this Sam framework is, or I would say internally, it's actually calling a cloud formation template. So what I'm saying to them is just create an API gateway for me and just also share the API gateway URL at the end of it. After deploying the same template. It's going to generate a URL. The format of URL will be something like this. This will be kind of a name of API gateway execute API is going to make can execute call to this API. We have this prod which will be the default stage environment. And when I'm going to hit this URL again, it's a gate request call. So it's going to make a call to that lambda and it's going to print hello world. So this hello world is same like which we have seen on a local system when the code was running on locally. This is the same thing over here, and this information is coming from the router. So we have seen that we can do kind of code sharing part of that thing. We can kind of write the same code which can be used by the lambda and local. Now this is mostly all about the AWS lambda. Now let's come back to the next section, which is our graphQl. So GQL gen. So I'm going to most talk about the GraphQL part. So GraphQL is a query language for Apan that was developed in 2012 by Facebook and they made open source as 2015. It was mostly designed to improve the efficiency and flexibility of API by requesting to fetch the data which is only required, and to retrieve multiple set of data in a single request. And GraphQL is a strong type schema that offers structure of data and that can query and query language that allows clients to specify the data they want to retrieve. It can be used in any programming language and backend technology. So when you are talking about the graphql, we only need to think about the schema. So this schema will define everything. It defines what kind of model we want, how we want to tech that model, and how we want to update that model. So schema is kind of a core for graphQl. And after that there are two important concepts, which is called as Qian mutation. In case of traditional rest application, we have multiple things to get to update the data. So for example, we have put request, post request, delete request. So these are all the operations which is used to kind of update the data. In case of GraphQL, that is called as mutation. So whenever you want to perform any sort of update, which also include the deletion of that object, it's called as mutation. And when we want to tech the data that is called as query. So q is nothing but just get the data and mutations, just kind of modify the data. Let me show you a quick example of GraphQL. Schema. So we have query for tech the data, we have mutation for update the data and we have different, different models. So now let's consider a simple example that we want to create a to do app. The to do app required an id title description and completed to get the data from the API. We are going to have another input. When we are saying that 4k if you want to create to do, then we need to pass a title and description. Here the description is optional. If you see an exclamatory mark that is, this is required property. Similarly, when you want to update any of the to do, we need to pass the title description and the status whether it's completed or not. Okay, in case when you want to query here we say that, okay, we are going to get the to DOS and to DOS will going to get the list. So if you can see this is array option, we have square bracket inside. We are passing a model, the model is required and this return is also required. Even if it's going to return the empty object, that is perfectly fine. Similarly, mutation, what are the options we can perform? We can perform create to do, update to do and delete to do. Now this is all about the Graphql schema. Now what happened is now someone is going to say, okay, if I want to get the data, if I want to update the data. In our typical restful application we used to create a controller. We used to create a routing and everything. In case of graphql, instead of creating a controller router, we need to kind of mention the resolver for example. Now this is kind of a simple example, let's say if I want to get the to do which has id title completed. So this information will be linked to a particular set of function which is Golang to execute. Now it may be possible that the id title and completed these are going to come from different, different functions. So we can have a nested resolver. Also in our graphql GQl generation we will see that how these things are getting linked. So please bear with me for another five minutes. Now, what is GQL gen? So Gql Gen is a go library for building a graphql API. It generates typescript server code based on the GraphQL schema and resolver functions that you define. So we talk about the schema. The resolver is nothing but the function which is getting execute and is going to either update the data or is going to get the data. So with GQL you can define your graphQl schema in a graphQl schema language and graphQl generates go code for your server that handles the queries and mutations. This eliminates the need for manually parsing incoming requests and serializing outgoing responses. The GQL is built with a performance in mind and uses generation to code efficient and types of code. To getting started with the GQL gen, we need to run a simple command. This command is nothing good. To initialize the GQL gen, we need to run the command go, run GitHub.com 99 designs and GQL gen initialize. It's going to initialize the setup. It's going to create a dummy project for you. And once the setup is done, we also need to resolve the dependency. I already did this thing, so let me quickly walk you through on the code part, what it has been generated, and then we'll go in more details about how we are going to integrate with AWS lambda. After running GQL initialize command is going to create a folder structure which is equivalent to this one. The starting part for the GQL is to understand the GQL gen yaml file. And this YaMl file has lot of information. So let's go by a bit of information from here. So it checks with a schema where the schema is available. So here it's mentioned that the schema is available on a graphql folder. So it go to the Graphql folder, it will check for this file extension, Graphql S. And here it's going to find out the schema. And based on this schema, it's going to create the resolver and the model, and then it's going to check what will be the file path. If you want to change the file path, we can change the file path. Here it has uses this graphql generator go. So this is the generator Go is a kind of auto generated file after the initialization. Similarly, it's going to create a models for us. The model will be, if you can go back to our schema, we'll see that we have this to do. This is one model and new to do. This is another model. So it's going to create models new to do and to do based on the file which you are going to provide. So here it's mentioned that the models underscore gen Go is the place where it's going to create a model. Now if you already have can existing model and you don't want to create those models by this code, gen by Jigger gen. So what you can do is you can have one more option which is called as autobind. And here you can mention the path of your model. For example, in one of the example I'm going to show you that I already have a path where I mentioned my model and then I'm just telling the GQ gen that use the existing model, don't create the new model in models underscore gen file. So this queries and mutations, it has different, different kind of functionality. For example, we require one resolver. Resolver is nothing but the function when someone is going to request for this query and then we require another create to do. I just want to kind of create as simple as possible. So I just put the create to do where I'm going to put the to do information and it's golang to create this thing. So the base class for the resolver is resolver go. It has nothing, it has just a simple thing. This can be used as a dependency injections. I'm going to cover that part as well and then it's going to create a schema now. So if you want to have a separate schema based on the file name, that is also possible. That is something that we can configure here, that what kind of schema we want. And it's mentioned that, okay, just follow the schema based on the schema, just create the resolver. So we have this schema resolver and it has multiple function which is actually not implemented. For example, we are looking for to DOS to get the kind of curie information and it's mentioned here that this to do does not implement it. This is where we need to put our code. Similarly, it has created a resolver for create to do, which is mentioned here. So we require one more resolve for the create to do and it has created a resolver that is also not implemented. So we need to put the logic over here to create it. Now thing is, I just want to show you the entire end to end flow where I can perform a basic code operation for that thing. And I'm very fond of AWS. So what I've done so far is I created one AWS sample where it's going to kind of get the data from the dynamodb and update the data from the dynamodb. But before Golang to more on how this is going to be implemented, I just created a very simple sample where we can see that how we can perform basic code operation on DynamoDB. The main thing is with the dynamodb is code is a bit complex to understand, but there are a lot of libraries which are kind of a wrap, which is kind of a wrapper on the dynamodb to make our life easy. And one of them is dynamo. Now what this dynamo does is we need to mention the schema. So it's like we have can object where we need to just do a card operation, basic create, update, delete operation. And we need to say, okay, now this id is linked to particular this id into the table dynamodb table. Similarly, user id will link to user id table. And likewise we have other properties which can link to specific attribute to the dynamodb table. So consider this, because Dynamodb is like a key value pair database. So we can't use a term column, but right now we can use a term kind of column where we have this id, user id text and done. These are kind of attributes or we can say properties of a particular document in Dynamodb. And then we're just initializing a new AWS session. So this new session is in case when we want to kind of do anything related to the AWS, we need to initialize this AWS session here. We can also provide which region we want. We can also provide the credentials AWS. Well, I'm going to pass the credentials and region using an environment variable which I'm going to show you in a few minutes. And then I'm just initializing my dynamo which is kind of a wrapper on the dynamodb library provided by AWS. And then I'm just selecting a database table. For me, I just already created a table to save all of our time. And then I'm just doing creating an object for the table. It's like a struct for the table, running a put command, it's going to add this information. And then here I change a schema a bit, I'm passing a user id with this thing and at the end I'm just making a call to again the dynamodb table here. I'm saying, okay, give me the information based on the user id and it's just returning me this result and this result will be visible. A very simple example. So the main purpose of showing this example is that for our actual application we are going to use dynamodb where we are going to see the things end to end. Okay, now the GQL code which has been generated so far, I modified that code. So this is kind of updated code so far. It's a bit clean code compared to what has been generated. So let me quickly walk you through on that code. First here I just mentioned AWS where client I'm just provides a session information. Okay, this is the session information for us. We have this graphql which is same auto generated. Now we have a model to do model which I shown you before. We have this id, user id text dynamite. These are the address on that part. Now I'm just mentioning this GQL generate that, do not create a new model for moss, don't create a new model for me. So it's not created over here. And this is possible by providing a path where it can use an autobanding. So my to do models, this model folder is a part of this autobanding. So GQlgen will say okay, this model already exists, so I don't need to create another one. And then that's why it's just ignored that model. Then I created a service because the code which was in dynamo, it's just for the PoC, not for the actual code. So I just created a to do service. And this to do service is actually doing two things here. It's just doing adding a to do. In that case it's just generating a unique id, it's making a dynamodb table call, it's making a put request, getting the data and then we are good to go. The new table will be added when someone is going to call to this service. And similarly we need to fetch all the users informations. We need to fetch all the 2d information by the user. So here I'm passing a user id, there it is again, scanning the call, it's passing the user id and then we are getting the results which is kind of modeled for us. So this is a simple service that we have now coming back to the server go code. So this code has been kind of modified a bit so that we should use the gin framework. So far the auto generated code from GqlGen does not have this concept called as gin. They have a documentation that how we can integrate with the gin. And I follow the same documentation and this is the code we have. Okay, so again the code start from very simple thing, main here we are mentioning the port number. If the port number is not mentioned, we are going with the default port number. We are initializing the gin router. We are actually initializing two more endpoints. One is query where we are going to perform the graphql operation and the playground. So this juque and gen support a playground. We can use the same playground again or we can use something else. I'm going to show you one more chrome extension that you can use and you can play around with the playground. So this graphql handler here, I'm just initializing the new dynamo client, initializing the new to do service and I'm passing this information to default handler server. So this handler server is actually provides by our GQLGen, it's provided by GQGen and I'm just initializing this handler server and then when the server is initialized I'm passing this information to the gin context for the resolver, the resolver which has been generated here I'm passing the to do service as a dependency over here so that the resolver have all the information about the to do service. So this is the basic configuration that we have done so far and now the GQL gen has know about to do service and it has the instance of for the to do service. Now if I'm going back to my schema again now I need to fetch the data for the to Dos and I need to create a new to do for that. The GraphQL GQ engine has already created a resolver for us and here we are fulfilling that information. So for the create to do which is for the mutation operation we are just adding into to do which include the text user id and done current status which is false. And for fetching all the to DOS we are just passing the user id and it will return us information over all the user ids. Sorry it will return the information about all the to DOS based on the user id. To quickly see this in action I'm going to run here. So let me stop my existing server, let me go to specific this folder and here like I said that I'm going to run this information on AWS cloud and I'm going to use my existing profile. So you can pass your AWS credential here or you can also configure AWS provides for the same I'm just configuring the AWS profile and thanks to take nine so they are allowing me to use their AWS cloud for this purpose and I'm mentioning the region information. So these two information is required for dynamodb and then I can run the same command go run server. Now what can I do here is I can quickly show you that the entire thing in a now it has started the server. Now we can see that it has two endpoints. One is a gate request endpoints which is pointing to a playground and we have post call which is actually calling to this which is actually handling this resolver schema and everything. So when we open this port number 8080 where the gqlgen server is running. It is going to show us the playground. So this is kind of playground. So we can see there are multiple program provided. This is default playground provided by the gqueengen. And we can also see the information about the schema. So if you can see the documents, it has cure and mutation. The query has this to do where we need to pass the user id and mutation where we can kind of create a new to do. So I'm quickly going to show you a simple example. We have query to do if you can see to do, it's asking for user id and then text and user id. So this is the information when we are saying that, okay, this is the information we are passing you. And at the written just return me text and user id. If I'm going to run this thing. We can see we are getting text and user id. If I'm going to bypass any of the field, let's say I only need a text because I already know the user id this information is going to become. Now we can see that the information which is coming from backend is something that we are asking for. So this is one of the kind of a major advantage of graphQl. Now apart from query, we can also run mutation where we can create a new data. So here we are saying that, okay, create a to do, just pass some random values over here, the text and for the user id. So if I'm Golang to say for the user id to create it and then return me only the id, if I'm going to run the same command here and say that I need to have for the second user get the data and I'm going to run this thing again. So you can see that it's just returning me for the user two and this is for the user one. So this is all about the playground in this thing. Now let's see how we can deploy the same thing on AWS lambda. Now for the lambda, what I've done so far is I just created a common code, which is mostly a router code, into the router go file. We can quickly see that we have this router. We have the information about the GraphqL handler, legal information. It just initialized the router and it just kind of returned the path of the router. And then I created two different folders, one for local, for the local which is for the local operation, the offline mode, and then lambda main go, which is to run the application on lambda the main local main go is nothing but a simple initializing router, passing the port number to the router and then lambda has the same configuration, initializing a router which you have seen as a code sharing slide. And then we are just creating a proxy and we are just starting the lambda. So that's it. We are doing it for the deployment. The deployment is almost same, nothing has been changed. Only one thing which has been changed so far in the deployment is because this application require a dynamodB access. So we actually added a dynamodb access over here. We are just passing a customized role information. We are saying it can perform gate port queries, can update operation on the DynamodB database. And then the function that we have, I think everything is same. The path is lambda folder. We have kind of customized name for the handler name. It is using this runtime go one x version and it is using a specific role and that role has been created over there which has the dynamic access. Rest of the things are same. Now one more, there is a small, very small difference over there. If you can see the queries, it has only two things. So we have this gate query for the playground and for the post query to do kind of a query or mutation. So in graphql everything is mostly a post operation. So if you want to update something we need to use the HTTP work which is post or if you want to use any, get the data. We again go in for the post operation. So this is the entire configuration that we have here. We are mentioning that we need to kind of open two endpoint. One will be a gate and another will be a post. Post is for QD and gate is for the playground and that's it. And then again we need to run the same command, same build for building this structure and same deploy. And during the same deploy it is going to ask you a lot of things like what will be a cloudformation stack name, other information, your bucket name. And once the deployment is done you are going to get a URL. And this time I'm going to show you a different chrome extension which is for the playground. And we'll see how we can play around with the same thing. And the code is deployed on AWS lambda. So this is kind of a different variation of playground. So we have seen this playground which is provided by the GQL gen and this is kind of a chrome extension for us which is kind of very standard playground for graphQl operation. We can see schema here, we can see the schema over here and then we can perform the same operation. We can do again the query call with gate and the user information is going to return the data. We can quickly check for different user is returning this data. I can check for the specific thing and the important thing is the URL which you are seeing over here. This is actually pointing to an API gateway and this is the endpoint where we are heading so far. Similarly we can also do the mutation here. I can put some more content here, update and then again run the same query for the same user. And now you can see my contain this value which has been put it's available. So this is all about how we can use the GQN generator for creating the resolvers and how we can kind of deploy this thing on AWS lambda. Thank you so much for watching this presentation. I hope that you have learn lot of things from these presentations. And before I sign off I just want to give some information about my current company. I'm working in take nine and this is the fastest growing company in the world. We have offices in US, India and Latin America and you can reach out us on LinkedIn as well as you can call us on the mentioned mobile number. So thank you so much.