add

aws/proposals/dice-cloudwatch-datadog/cloudwatch-alarms.md

@@ -0,0 +1,91 @@
+# Cloudwatch Alarms to Datadog Monitors
+
+```yaml
+monitor:
+  - service: ECS
+    description: CPU Utilization
+    aws_metric:
+    datadog_metric:
+    groups:
+      - name: dice-mojo-common
+        filter: dice-mojo
+        tags: 
+          environment: common
+          boc_program: dice
+        warning:
+          threshold:
+        critical:
+          threshold:
+
+```
+
+# notification teams
+
+For each set of notification target groups, create a Datadog team.
+
+dice-{project}-{env}[-{service}]
+
+* {project}
+  * mojo
+  * centurion
+  * auth
+  * cumulus
+
+* {env}
+  * common
+  * prod
+  * dmz-stage
+  * dmz-prod
+
+* {service}
+  * where necessary to have a different group, use a short service name
+  * rds, for example, goes to a different group of people
+
+# sns
+
+For the same set of notification teams, create an SNS topic.
+In that topic, subscribe ONLY the refactored lambda.
+
+The lambda will read the SNS message, determine who should be notified
+from an SSM parameter
+
+# services
+
+* ecs
+* lambda
+* alb
+* rds
+
+# monitors
+
+For each service and metric, create a monitor.  Each monitor has two thresholds,
+warning and alert (critical).
+
+Each monitor will have a templated alert message.  The warning will indicate
+a warning of VALUE > THRESHOLD, and will include the specific team plus any service
+specific teams (if they exist).
+
+The critical will indicate the same detail, will send to the same DD teams,
+as well as to the SNS topic.
+
+# 
+
+# ssm parameter
+
+/apps/dice/datadog/notifications
+  {notification-group}   # same as dd team
+
+```yaml
+name: {notification-group}
+alerts:
+  warn: false
+  critical: true
+  recovery: true
+contacts:
+  - username: jbid
+    mail: email-address
+    telephonenumber: +1-xxx-xxx-xxxx
+  - username:
+    mail: 
+    telephonenumber:
+```