Add healthmonitor to Octavia

Healthmonitor is Octavia's main way to check on the status of the
octavia backend members, which makes it an important part of Octavia.

After any disruptive operation, if any of the backend member servers
would get an error, then without the healthmonitor it will be harder
to spot that, as the octavia member servers will hold the NO_MONITOR
operating_status, and traffic to the LB might fail.

With the healthmonitor, the octavia member servers will enter an
ERROR operating_status, and so will the pool and the LB.
Also, we could recognize the problem we had with the members in an
easier way.

This patch adds the octavia healthmonitor to Octavia's pool stack.

Change-Id: I29c7888e04bcc1fe4cedfdfb572cbf657f5e711a

tobiko/openstack/stacks/_octavia.py

@@ -125,6 +125,15 @@ class OctaviaPoolStackFixture(heat.HeatStackFixture):
 
     hm_type = 'HTTP'
 
+    # healthmonitor attributes
+    hm_delay = 3
+
+    hm_max_retries = 4
+
+    hm_timeout = 3
+
+    hm_type = 'HTTP'
+
     @property
     def listener_id(self):
         return self.listener.listener_id

tobiko/openstack/stacks/octavia/pool.yaml

@@ -13,15 +13,36 @@ parameters:
         default: HTTP
         description: Protocol used by the pool members
 
+    listener_id:
+        type: string
+        description: ID of the listener
+
+    # healthmonitor params
+    hm_delay:
+        type: number
+        default: 10
+        description: >
+            The minimum time in seconds between regular connections of the
+            member.
+    hm_max_retries:
+        type: number
+        default: 4
+        description: >
+            Number of permissible connection failures before changing the member
+            status to INACTIVE.
+
+    hm_timeout:
+        type: number
+        default: 3
+        description: >
+            Maximum number of seconds for a monitor to wait for a connection
+            to be established before it times out.
+
     hm_type:
         type: string
         default: HTTP
         description: Type of health-monitor
 
-    listener_id:
-        type: string
-        description: ID of the listener
-
 resources:
     pool:
         type: OS::Octavia::Pool
@@ -30,7 +51,20 @@ resources:
             protocol: { get_param: pool_protocol }
             listener: { get_param: listener_id }
 
+    monitor:
+        type: OS::Octavia::HealthMonitor
+        properties:
+            delay: { get_param: hm_delay }
+            type: { get_param: hm_type }
+            timeout: { get_param: hm_timeout }
+            max_retries: { get_param: hm_max_retries }
+            pool: { get_resource: pool }
+
 outputs:
     pool_id:
         description: Pool ID
         value: { get_resource: pool }
+
+    monitor_id:
+        description: Healthmonitor ID
+        value: { get_resource: monitor }