Source: faun.pub
Why We Moved From Lambda to ECSCategory: Database, Data, Redis, Terraform
After many months of development, my team just announced the general availability of our platform. In this post, I’ll dive into what went well in Lambda, what challenges we faced, and why we eventually made the decision to migrate some services from Lambda to AWS Elastic Container Service (ECS).
We had a number of issues, ranging from speed to SQS size limits and lack of process isolation in Lambda, that caused us to reconsider its effectiveness as our integration runner.
We had to reinvent some wheels that Lambda had given us for free - like logging, autoscaling, and health checks - but in the end, we had our 6-step test integration back to running in under 2 seconds.
For our integration runner, though, we learned that the size, speed, and process isolation limitations of Lambda made ECS a better option, and it was worth the dev time it took to create an ECS deployment for that particular service.
We had a number of issues, ranging from speed to SQS size limits and lack of process isolation in Lambda, that caused us to reconsider its effectiveness as our integration runner.
We had to reinvent some wheels that Lambda had given us for free - like logging, autoscaling, and health checks - but in the end, we had our 6-step test integration back to running in under 2 seconds.
For our integration runner, though, we learned that the size, speed, and process isolation limitations of Lambda made ECS a better option, and it was worth the dev time it took to create an ECS deployment for that particular service.
Related Articles
Community Partners
DevOps Careers