After upgrading to release v7.11.0, we are getting the following error when Terraform tries to create the Lambda event source mapping for the termination watcher:
Error: creating Lambda Event Source Mapping
(arn:aws:sqs:eu-north-1:058264392402:gha-ghec-eu-deregister-retry):
operation error Lambda: CreateEventSourceMapping,
https response error StatusCode: 400,
RequestID: 2fbd72b3-9a40-4ed8-aa7e-00daa25fdfd2,InvalidParameterValueException:
Queue visibility timeout: 60 seconds is less than Function timeout: 300 seconds
with module.gha_ghec_eu.module.github_runners.module.instance_termination_watcher[0]
.aws_lambda_event_source_mapping.deregister_retry[0],on .terraform/modules/gha_ghec_eu.github_runners/modules/termination-watcher/deregister-retry.tf line 54,
in resource "aws_lambda_event_source_mapping" "deregister_retry":
54: resource "aws_lambda_event_source_mapping" "deregister_retry" {
We have enabled instance_termination_watcher with a Lambda timeout of 300 seconds:
instance_termination_watcher = {
enable = true
enable_metric = {
spot_warning = true
}
timeout = 300
zip = data.artifactory_file.termination-watcher.output_path
}
The issue appears to be related to the SQS visibility timeout configured in: https://github.com/github-aws-runners/terraform-aws-github-runner/blob/v7.11.0/modules/termination-watcher/deregister-retry.tf#L23
Could you please confirm whether this is an issue in v7.11.0 and whether the SQS visibility timeout should be derived from the configured Lambda timeout?
After upgrading to release v7.11.0, we are getting the following error when Terraform tries to create the Lambda event source mapping for the termination watcher:
We have enabled instance_termination_watcher with a Lambda timeout of 300 seconds:
The issue appears to be related to the SQS visibility timeout configured in: https://github.com/github-aws-runners/terraform-aws-github-runner/blob/v7.11.0/modules/termination-watcher/deregister-retry.tf#L23
Could you please confirm whether this is an issue in v7.11.0 and whether the SQS visibility timeout should be derived from the configured Lambda timeout?