Smoke test: also test deploy alerting module with terraform (#1029)

* Smoke test include en-alerting module

* Terraform delete more resilient

* Depends on module.en

scripts/ci-terraform.sh

@@ -138,7 +138,7 @@ function destroy() {
   # especially when previous terraform deployment failed. So grepping from terraform state instead.
   db_inst_name="$(terraform state show module.en.google_sql_database_instance.db-inst | grep -Eo 'name.*=.*"terraform-[a-zA-Z0-9]+"' | grep -Eo 'terraform-[a-zA-Z0-9]+' | uniq)" || best_effort
   if [[ -n "${db_inst_name}" ]]; then
-    gcloud sql instances delete ${db_inst_name} -q --project=${PROJECT_ID}
+    gcloud sql instances delete ${db_inst_name} -q --project=${PROJECT_ID} || best_effort
   fi
   # Clean up states after manual DB delete
   terraform state rm module.en.google_sql_user.user || best_effort
@@ -149,6 +149,7 @@ function destroy() {
   local vpc_connector_name="serverless-vpc-connector"
   gcloud compute networks vpc-access connectors delete ${vpc_connector_name} -q --region us-central1 --project=$PROJECT_ID || best_effort
 
+  echo "Start terraform destroy"
   terraform destroy -auto-approve
   popd > /dev/null
 }

terraform-e2e-ci/main.tf

@@ -55,6 +55,19 @@ module "en" {
   }
 }
 
+module "en-alerting" {
+  source = "../terraform/alerting"
+
+  monitoring-host-project     = var.project
+  verification-server-project = var.project
+
+  adminapi_hosts  = ["adminapi.notexistdomain.notexist"]
+  apiserver_hosts = ["apiserver.notexistdomain.notexist"]
+  server_hosts    = ["encv-test.notexistdomain.notexist"]
+
+  depends_on = [module.en]
+}
+
 output "en" {
   value = module.en
 }