Introduction

If you've been battling unexpected request timeouts in your AWS ECS Service Connect setup, you're not alone. I recently ran into an issue where long-running processes mysteriously failed—only to discover that the root cause was the default 15-second request timeout for HTTP and gRPC requests. 😱

Here’s the good news: As of January 2024, AWS now allows you to configure these timeout values. Let's explore how.

Understanding the Default Timeout Behavior

Before this update, AWS ECS Service Connect enforced a fixed 15-second timeout on each request. If your process took longer—boom—it was terminated. This default wasn’t exposed in your service’s configuration, making it tricky to diagnose.

This was particularly painful for:

• Long-polling APIs

• Background job processors

• Real-time data pipelines

🆕 The January 2024 Update: Configurable Timeouts

In response to user feedback, AWS introduced configurable timeout settings for ECS Service Connect. 🎉

You can now define:

• **idleTimeoutSeconds**: How long to keep idle connections open

• **perRequestTimeoutSeconds**: Maximum duration for a single request

How to Update Timeout Settings

Prerequisites

• Updated AWS CLI (post-January 2024 version)

• Task definitions with portName values

Timeout Parameters Explained

ParameterDefaultPurposeidleTimeoutSeconds300 (TCP), 3600 (HTTP)Max idle time before closing a connectionperRequestTimeoutSeconds15Max time allowed for a single request

Modify Your ECS Service Definition

Here's a sample update to your service definition JSON:

{

"serviceConnectConfiguration": {

"services": [

{

"clientAliases": [

{

"port": 3000

}

],

"portName": "your-port-name",

"timeout": {

"idleTimeoutSeconds": 3600,

"perRequestTimeoutSeconds": 3600

}

}

]

}

}

You can apply this via:

• AWS CLI

• CloudFormation

• Tools like ecspresso

✅ Verifying the Configuration

After applying changes, verify them with:

aws ecs describe-services \

--cluster your-cluster-name \

--services your-service-name

📝 Note:
Default values like the 15-second request timeout won’t appear unless explicitly defined—so always set them manually if needed.

⚠️ Important Considerations

• No Documented Upper Limit
  AWS hasn’t shared a max value for timeouts—set them wisely based on use case.

• Reevaluate Architecture for Long Tasks
  Long timeouts may not always be ideal. Consider:

  • Async job processing

  • Decoupled service patterns

  • Background queues (e.g., SQS, SNS, Step Functions)

Conclusion

The 15-second default timeout in ECS Service Connect can quietly break your long-running workflows. But with AWS’s new configurable timeout options, you can take control—tailoring connection behavior to your app’s needs.

Just remember:

• Explicitly define perRequestTimeoutSeconds and idleTimeoutSeconds

• Keep your CLI and SDK updated

• Validate your settings after deployment

With these tweaks, your services will run smoother, longer, and without those unexpected timeouts.