Granting permissions

Allow Lambda to read/write our DynamoDB table #

Let’s give our Lambda’s execution role permissions to read/write from our table.

Go back to src/CdkWorkshop/HitCounter.cs and add the following highlighted lines:

using Amazon.CDK;
using Amazon.CDK.AWS.Lambda;
using Amazon.CDK.AWS.DynamoDB;
using Constructs;
using System.Collections.Generic;

namespace CdkWorkshop
{
    public class HitCounterProps
    {
        // The function for which we want to count url hits
        public IFunction Downstream { get; set; }
    }

    public class HitCounter : Construct
    {
        public IFunction Handler { get; };

        public HitCounter(Construct scope, string id, HitCounterProps props) : base(scope, id)
        {
            var table = new Table(this, "Hits", new TableProps
            {
                PartitionKey = new Attribute
                {
                    Name = "path",
                    Type = AttributeType.STRING
                }
            });

            Handler = new Function(this, "HitCounterHandler", new FunctionProps
            {
                Runtime = Runtime.NODEJS_14_X,
                Handler = "hitcounter.handler",
                Code = Code.FromAsset("lambda"),
                Environment = new Dictionary<string, string>
                {
                    ["DOWNSTREAM_FUNCTION_NAME"] = props.Downstream.FunctionName,
                    ["HITS_TABLE_NAME"] = table.TableName
                }
            });

            // Grant the lambda role read/write permissions to our table
            table.GrantReadWriteData(Handler);
        }
    }
}

Deploy #

Save & deploy:

cdk deploy

Test again #

Okay, deployment is complete. Let’s run our test again (either use curl or your web browser):

curl -i https://xxxxxxxxxx.execute-api.us-east-1.amazonaws.com/prod/

Again?

HTTP/1.1 502 Bad Gateway
...

{"message": "Internal server error"}

😢 #

Still getting this pesky 5xx error! Let’s look at our CloudWatch logs again (click “Refresh”):

{
    "errorMessage": "User: arn:aws:sts::585695036304:assumed-role/CdkWorkshopStack-HelloHitCounterHitCounterHandlerS-TU5M09L1UBID/CdkWorkshopStack-HelloHitCounterHitCounterHandlerD-144HVUNEWRWEO is not authorized to perform: lambda:InvokeFunction on resource: arn:aws:lambda:us-east-1:585695036304:function:CdkWorkshopStack-HelloHandler2E4FBA4D-149MVAO4969O7",
    "errorType": "AccessDeniedException",
    "stackTrace": [
        "Object.extractError (/var/runtime/node_modules/aws-sdk/lib/protocol/json.js:48:27)",
        "Request.extractError (/var/runtime/node_modules/aws-sdk/lib/protocol/rest_json.js:52:8)",
        "Request.callListeners (/var/runtime/node_modules/aws-sdk/lib/sequential_executor.js:105:20)",
        "Request.emit (/var/runtime/node_modules/aws-sdk/lib/sequential_executor.js:77:10)",
        "Request.emit (/var/runtime/node_modules/aws-sdk/lib/request.js:683:14)",
        "Request.transition (/var/runtime/node_modules/aws-sdk/lib/request.js:22:10)",
        "AcceptorStateMachine.runTo (/var/runtime/node_modules/aws-sdk/lib/state_machine.js:14:12)",
        "/var/runtime/node_modules/aws-sdk/lib/state_machine.js:26:10",
        "Request.<anonymous> (/var/runtime/node_modules/aws-sdk/lib/request.js:38:9)",
        "Request.<anonymous> (/var/runtime/node_modules/aws-sdk/lib/request.js:685:12)"
    ]
}

Another access denied, but this time, if you take a close look:

User: <VERY-LONG-STRING> is not authorized to perform: lambda:InvokeFunction on resource: <VERY-LONG-STRING>"

So it seems like our hit counter actually managed to write to the database. We can confirm by going to the DynamoDB Console:

But, we must also give our hit counter permissions to invoke the downstream lambda function.

Grant invoke permissions #

Add the highlighted lines to src/CdkWorkshop/HitCounter.cs:

using Amazon.CDK;
using Amazon.CDK.AWS.Lambda;
using Amazon.CDK.AWS.DynamoDB;
using Constructs;
using System.Collections.Generic;

namespace CdkWorkshop
{
    public class HitCounterProps
    {
        // The function for which we want to count url hits
        public IFunction Downstream { get; set; }
    }

    public class HitCounter : Construct
    {
        public IFunction Handler { get; };

        public HitCounter(Construct scope, string id, HitCounterProps props) : base(scope, id)
        {
            var table = new Table(this, "Hits", new TableProps
            {
                PartitionKey = new Attribute
                {
                    Name = "path",
                    Type = AttributeType.STRING
                }
            });

            Handler = new Function(this, "HitCounterHandler", new FunctionProps
            {
                Runtime = Runtime.NODEJS_14_X,
                Handler = "hitcounter.handler",
                Code = Code.FromAsset("lambda"),
                Environment = new Dictionary<string, string>
                {
                    ["DOWNSTREAM_FUNCTION_NAME"] = props.Downstream.FunctionName,
                    ["HITS_TABLE_NAME"] = table.TableName
                }
            });

            // Grant the lambda role read/write permissions to our table
            table.GrantReadWriteData(Handler);

            // Grant the lambda role invoke permissions to the downstream function
            props.Downstream.GrantInvoke(Handler);
        }
    }
}

Diff #

You can check what this did using cdk diff:

cdk diff

The Resource section should look something like this, which shows the IAM statement was added to the role:

Resources
[~] AWS::IAM::Policy HelloHitCounter/HitCounterHandler/ServiceRole/DefaultPolicy HelloHitCounterHitCounterHandlerServiceRoleDefaultPolicy1487A60A
 └─ [~] PolicyDocument
     └─ [~] .Statement:
         └─ @@ -19,5 +19,15 @@
            [ ]         "Arn"
            [ ]       ]
            [ ]     }
            [+]   },
            [+]   {
            [+]     "Action": "lambda:InvokeFunction",
            [+]     "Effect": "Allow",
            [+]     "Resource": {
            [+]       "Fn::GetAtt": [
            [+]         "HelloHandler2E4FBA4D",
            [+]         "Arn"
            [+]       ]
            [+]     }
            [ ]   }
            [ ] ]

Which is exactly what we wanted.

Deploy #

Okay… let’s give this another shot:

cdk deploy

Then hit your endpoint with curl or with your web browser:

curl -i https://xxxxxxxxxx.execute-api.us-east-1.amazonaws.com/prod/

Output should look like this:

HTTP/1.1 200 OK
...

Hello, CDK! You've hit /

If you still get 5xx, give it a few seconds and try again. Sometimes API Gateway takes a little bit to “flip” the endpoint to use the new deployment.

😲 #

We use analytics to make this content better, but only with your permission.

More information