Snowplow on AWS Fargate - Task Role

February 10, 2020

This is part two of a blog post series about Snowplow on AWS Fargate.


This post will outline the problem of using the IAM role/policy granted to AWS Fargate tasks so Snowplow components can access AWS resources.

Using permissions in ECS task role

The first problem you’re likely to run into is permission errors related to assuming the task role in your AWS ECS task definition. Lucky for us, this is just a minor configuration tweak.

If you follow along with the documentation for configuring the scala stream collector and use start with the default application.conf you’ll run into this problem.

By default, the application.conf file uses the value iam for its aws.accessKey and aws.secretKey configuration values. This value works for deployment directly on EC2 instances but does not work when running as an ECS Fargate task. For these, you need to use the AWS default credential provider. Thankfully, Snowplow supports this, so just a simple change from iam to default will get you around this problem.

aws {
 -    accessKey = iam
 +    accessKey = default
 -    secretKey = iam
 +    secretKey = default

If you’re following along with the Snowplow Wiki setup guide, you’ll first run into this with the Scala Streaming Collector. However, this change will be required for all components being run as Fargate tasks, e.g. Stream Enrich, S3 Loader, ES Loader, etc.


Check out the next post in this series, Snowplow on AWS Fargate - Stream Enrich, which covers the Stream Enrich process on AWS Fargate.

Return home