set up metric filter pattern and fix tf apply

Author: Akol125

infrastructure/instance/endpoints.tf

@@ -15,6 +15,7 @@ module "get_status" {
   function_name = "get_status"
   image_uri     = module.docker_image.image_uri
   policy_json   = data.aws_iam_policy_document.logs_policy_document.json
+  aws_sns_topic = data.aws_sns_topic.fhir_api_errors.arn
 }
 
 locals {

infrastructure/instance/modules/lambda/lambda.tf

@@ -55,8 +55,8 @@ resource "aws_cloudwatch_log_metric_filter" "fhir_api_error_logs" {
   count = var.error_alarm_notifications_enabled ? 1 : 0
 
   name           = "${local.short_prefix}_${var.function_name}-ErrorLogsFilter"
-  pattern        = "%\\[ERROR\\]%"
-  log_group_name = aws_cloudwatch_log_group.api_access_log.name
+  pattern        = "{ $.operation_outcome.status = \"500\" || $.operation_outcome.status = \"403\" }"
+  log_group_name = module.lambda_function_container_image.lambda_cloudwatch_log_group_name
 
   metric_transformation {
     name      = "${local.short_prefix}_${var.function_name}-ApiErrorLogs"
@@ -77,6 +77,6 @@ resource "aws_cloudwatch_metric_alarm" "fhir_api_error_alarm" {
   statistic           = "Sum"
   threshold           = 1
   alarm_description   = "This sets off an alarm for any error logs found in fhir api Lambda function"
-  alarm_actions       = [data.aws_sns_topic.fhir_api_errors.arn]
+  alarm_actions       = var.aws_sns_topic != null ? [var.aws_sns_topic] : []
   treat_missing_data  = "notBreaching"
 }

infrastructure/instance/modules/lambda/variables.tf

@@ -10,6 +10,16 @@ variable "function_name" {
   type = string
 }
 
+variable "aws_sns_topic" {
+  description = "SNS topic ARN for CloudWatch alarm notifications"
+  type        = string
+  default     = null
+}
+variable "error_alarm_notifications_enabled" {
+  description = "useful switching error alerting between environment"
+  type        = string
+}
+
 variable "image_uri" {
   type = string
 }