How to write task definition based on different enviroments - dev, staging & prod

Asked Jul 22 '23 at 19:57

Active Jul 22 '23 at 19:57

Viewed 22 times

I have an ECS service that is running a task with one container each. Challenge is here is that I have different compute configuration based on environment.

For example in dev environment I intend to use .25cpu with .5MB memory but in production I would want to provision it 2cpu and 4GB Memory.

My thought process for the above challenge:

Create separate task family for each service with separate configurations and container images. For ex: my-service-dev & my-service-prod
Create a standard template and sed service name to update values

What can be the most scalable way to tackle above challenge ?

asked Jul 22 '23 at 19:57

power-cut

1,310
1
14
30

i would use one pipeline and set a variable or secret with the environment name. – jmvcollaborator Jul 22 '23 at 20:01
@jmvcollaborator are you proposing to create task definition family based on the env like my-service-dev ? I don't if this would be scalable, as imagine of 10 services so 20 task definition family. – power-cut Jul 22 '23 at 20:10
i kinda envision a multistage unique pipeline – jmvcollaborator Jul 22 '23 at 20:15
@jmvcollaborator...hm. I dont know how scalable this would be. I can think of json merge using `yq merge` using json fragments ? – power-cut Jul 22 '23 at 20:18
I use Cloud Formation templates for this, and pass everything that varies by environment as parameters. – apokryfos Jul 23 '23 at 04:23

How to write task definition based on different enviroments - dev, staging & prod

0 Answers0