Aws_scheduler_schedule

Terraform Providers AWS
1

I am trying to use the aws_scheduler_schedule resource to stop /start an open search ingestion pipeline.

resource “aws_scheduler_schedule” “start” {
for_each = var.tenants
name = “{each.key}-pipeline-start" description = "Start schedule for {each.key} pipeline”
flexible_time_window {
mode = “OFF”
}
schedule_expression = “{each.key}.schedule_expression_start" target { input = jsonencode({ PipelineName= "alpha" }) arn = aws_osis_pipeline.this[each.key].***pipeline_arn*** # "aws_osis_pipeline.this[{each.key}.arn]”
role_arn = aws_iam_role.start_stop.arn
}
}

first issue ( not very important ) pipeline_arn is colored in red as if this attribute is not known
second issue when I try to create this resource I have the following error :
creating Amazon EventBridge Scheduler Schedule (search-pipeline-start): operation error Scheduler: CreateSchedule, https response error StatusCode: 400, RequestID: e96d7d24-ff5f-4b80-84b9-bd6cbc9d2fdb, ValidationException: osis is not a supported service for a target.

2

I’ve not really used EventBridge Scheduler, but looking at the documentation it seems that OSIS is not a supported templated target, so it falls under the universal target bucket. Thus the target ARN might be something like arn:aws:scheduler:::aws-sdk:osis:StartPipeline. You also need to provide input as a JSON since StartPipeline needs the pipeline name:

{
  "PipelineName": "MyPipeline"
}

Perhaps you can work backwards by first creating an schedule in the AWS Management Console, then look at the request parameters in the records CreateSchedule event in CloudTrail.

3

Might work?
Can you use reccurence? unsure to to translate it over. This just triggers a scaling at a certain time everyday. Will need to create an additional one to potentially stop/scale down what was created.

example:

resource "aws_autoscaling_schedule" "example" {
  scheduled_action_name  = "example"
  min_size               = 1
  max_size               = 3
  desired_capacity       = 3
  recurrence             = "0 11 * *"
  autoscaling_group_name = aws_autoscaling_group.example.name
}
4

Hi acwwat , you are right in fact it was the role that posed the problem , I didn’t use aws:scheduler:::aws-sdk:osis:StartPipeline . What I id is to create the schedule on the AWS and import it to find to right argumene . Here is the implementation ( I have a bunch of app thus the for each)
resource “aws_scheduler_schedule” “start” {
for_each = var.tenants
name = “{each.key}-pipeline-start" description = "Start schedule for {each.key} pipeline”
flexible_time_window {
mode = “OFF”
}
schedule_expression = each.value.schedule_expression_start
target {
input = jsonencode({
PipelineName = each.key
})
arn = “arn:aws:scheduler:::aws-sdk:osis:startPipeline”
role_arn = aws_iam_role.start_stop.arn

}
schedule_expression_timezone = var.schedule_expression_timezone

}