Schedule-meeting Lambda

When you create your first Lambda, the only thing you should ensure is that your runtime is Python. I went with Python 3.14.

Key Imports

The most important import for our Lambda is boto3. This allows us to perform read / write operations with other AWS services.

Critical Code Components

Here are the critical pieces of this Lambda:

Initial Setup

dynamodb = boto3.resource('dynamodb')
scheduler = boto3.client('scheduler')

TABLE_NAME = 'elephant-meetings'
table = dynamodb.Table(TABLE_NAME)
JITSI_DOMAIN = 'meet.christardy.com'

elephant-meetings is the name of my DynamoDB table
scheduler refers to EventBridge Scheduler. More on that in a second.
JITSI_DOMAIN refers to the domain that I selected to host my Jitsi server on.

Meeting ID and Jitsi URL Generation

meeting_id = generate_id()
jitsi_url = f"https://{JITSI_DOMAIN}/{meeting_id}"

Awesome - Jitsi Meeting Provisioning

Remember that Jitsi provisions meetings on the fly. The way I decided to do this is to just use the random primary key from DynamoDB as the meeting room name. That ensures that there won't be any collisions and I don't need to actually talk to the Jitsi service at all to get a meeting room.

EventBridge Scheduler

launch_dt = meet_dt - timedelta(minutes=10) 
launch_iso = launch_dt.strftime('%Y-%m-%dT%H:%M:%S')

print(f"Scheduling Launch for: {launch_iso}")

scheduler.create_schedule(
    Name=f"launch-{meeting_id}",
    ScheduleExpression=f"at({launch_iso})",
    Target={
        'Arn': launcher_arn,
        'RoleArn': scheduler_role,
        'Input': json.dumps({'meeting_id': meeting_id})
    },
    FlexibleTimeWindow={'Mode': 'OFF'},
    ActionAfterCompletion='DELETE'
)

Note - EventBridge Scheduler

This is a good time to talk about EventBridge Scheduler. It's a really intelligent AWS service that you can call directly from boto3. It has a GUI just like any other service in AWS, but we are instead calling it programmatically.

What it does is you can set a timed trigger independent of the Lambda, and have it execute specific behaviour. What I use it for is if a Lambda should call another Lambda, you can use EventBridge.

In this case, the payload from the frontend has a meeting time, which is when the teacher has scheduled the tutoring session. launch_dt is defined as that meeting time minus 10 minutes (EC2's need time to spin up, and we will be running our own health checks).

Here are a few more details about the EventBridge object we're creating:

Name: Assigns a unique identifier to this schedule using the meeting_id.
ScheduleExpression: Sets the exact execution time (at(...)) using an ISO timestamp.
Target: Defines what to trigger (the launcher_arn), the permissions to do so (RoleArn), and the JSON payload (Input) to send.
FlexibleTimeWindow: OFF ensures that Amazon executes it at the exact time specified.
ActionAfterCompletion: Automatically deletes the schedule entry after it runs once.

Testing - EventBridge Scheduler

Something that is useful for testing: if you schedule a meeting in the past, EventBridge scheduler will trigger right away.

Note - Safety Net Scheduler

There's another scheduler that handles the meeting safety net, which will be covered later in these docs, but in short, if no one shows up to the scheduled meeting after 15 minutes, it will kill the EC2 instances provisioned so that you don't rack up a huge AWS bill.

DynamoDB Payload

Here's the payload in schedule-meeting:

DynamoDB Payload

item = {
    'id': meeting_id, 
    'teacher_name': payload.get('teacher_name'),
    'student_name': payload.get('student_name'),
    'meet_time': meet_time_str, 
    'jitsi_url': jitsi_url,
    'status': 'SCHEDULED', 
    'created_at': datetime.now(timezone.utc).isoformat()
}

Note that this is the payload from the frontend, plus a few more params created in this function.

Lambda Response

return {
    'statusCode': 200,
    'headers': {
        'Content-Type': 'application/json',
        'Access-Control-Allow-Origin': '*',
        'Access-Control-Allow-Methods': 'OPTIONS,POST'
    },
    'body': json.dumps({
        'message': 'Session scheduled!', 
        'id': meeting_id,
        'jitsi_url': jitsi_url,
        'teacher_link': teacher_link, 
        'student_link': student_link 
    })
}

Once this function is successful, it's not just gonna send back "success" to the frontend. It's going to create the links for both the teacher and student to join their provisioned EC2 instances.
The frontend receives that statusCode 200 and will redirect to a modal providing the meeting links.

You'll notice that the links follow a specific format:

Link Format

teacher_link = f"{frontend_base_url}?id={meeting_id}&role=teacher"

Testing - Lambda Function

To test this Lambda, there is a test panel in the Lambda service. You can send a mock payload, which will actually write to your DB.

IAM Roles

You'll need a special IAM role for this Lambda. We previously set up IAM roles in Phase 2, so I'll speed through this.

Role 1: AWSLambdaBasicExecutionRole

Pre-defined by AWS

Role 2: SchedulerLogic

SchedulerLogic Policy

{
    "Version": "2012-10-17",
    "Statement": [
        {
            "Sid": "DatabaseWrite",
            "Effect": "Allow",
            "Action": "dynamodb:PutItem",
            "Resource": "arn:aws:dynamodb:*:*:table/elephant-meetings"
        },
        {
            "Sid": "CreateTimer",
            "Effect": "Allow",
            "Action": "scheduler:CreateSchedule",
            "Resource": "*"
        },
        {
            "Sid": "HandOverKeyCard",
            "Effect": "Allow",
            "Action": "iam:PassRole",
            "Resource": "arn:aws:iam::034489661489:role/ElephantSchedulerRole"
        }
    ]
}

Role 3: Write Meetings

Write Meetings Policy

{
    "Version": "2012-10-17",
    "Statement": [
        {
            "Sid": "VisualEditor0",
            "Effect": "Allow",
            "Action": "dynamodb:PutItem",
            "Resource": "arn:aws:dynamodb:*:034489661489:table/*"
        }
    ]
}

Environment Variables

Also, you're going to need a few environment variables:

Key
`LAUNCHER_ARN`
`SAFETY_NET_ARN`
`SCHEDULER_ROLE_ARN`

Note - Environment Variables

You can skip or leave the top two blank since we haven't written them yet. We're going to define a specific IAM role for the EventBridge scheduler within the schedule-meeting Lambda.

Warning - IAM Role Names

WARNING: they do NOT have the same roles, despite having similar names!

EventBridge Scheduler Role

Create a new IAM role, and apply the pre-defined AWS role called LambdaInvoke. If you don't see that, this is the policy:

LambdaInvoke Policy

{
    "Version": "2012-10-17",
    "Statement": [
        {
            "Effect": "Allow",
            "Action": "lambda:InvokeFunction",
            "Resource": "*"
        }
    ]
}

Copy the ARN in the top right of the IAM role and add it as an environment variable. This is what gives permissions to the scheduler to call another Lambda.

And that's the gist of the first Lambda, schedule-meeting. Now let's set up API Gateway so that we can hit it.

Next: API Gateway and Testing →