Document topicName and aggregation in gcp-pubsub

Signed-off-by: Kevin Mingtarja <[email protected]>

content/docs/2.13/scalers/gcp-pub-sub.md

@@ -16,17 +16,22 @@ triggers:
   metadata:
     subscriptionSize: "5" # Deprecated, use mode and value fields instead
     mode: "SubscriptionSize" # Optional - Default is SubscriptionSize - SubscriptionSize or OldestUnackedMessageAge
+    aggregation: "sum" # Optional - Only meaningful for distribution-valued metrics
     value: "5.5" # Optional - Default is 10
     activationValue: "10.5" # Optional - Default is 0
-    subscriptionName: "mysubscription" # Required
+    # Either one of subscriptionName or topicName is required
+    subscriptionName: "mysubscription"
+    topicName: "mytopic"
     credentialsFromEnv: GOOGLE_APPLICATION_CREDENTIALS_JSON # Required
 ```
 
-The Google Cloud Platform‎ (GCP) Pub/Sub trigger allows you to scale based on the number of messages or oldest unacked message age in your Pub/Sub subscription.
+The Google Cloud Platform‎ (GCP) Pub/Sub trigger allows you to scale based on any metrics from your Pub/Sub subscription or topic, such as number of messages or oldest unacked message age, etc.
 
 - `credentialsFromEnv` - This property maps to the name of an environment variable in the scale target (`scaleTargetRef`) that contains the service account credentials (JSON). KEDA will use those to connect to Google Cloud Platform and collect the required stack driver metrics in order to read the number of messages in the Pub/Sub subscription.
 
-- `mode` - The metric used to scale your workload. It's the camel case of the official metric name. For example, if you are going to leverage the metric `subscription/pull_request_count`, you will fill the value as `PullRequestCount`. Please refer to https://cloud.google.com/monitoring/api/metrics_gcp#gcp-pubsub. All metric starts with `subscription/` are supported. (Default: `SubscriptionSize`, aka `NumUndeliveredMessages`)
+- `mode` - The metric used to scale your workload. It's the `PascalCase` of the official metric name. For example, if you are going to leverage the metric `subscription/pull_request_count`, you will fill the value as `PullRequestCount`. Please refer to https://cloud.google.com/monitoring/api/metrics_gcp#gcp-pubsub. All metrics starting with `subscription/` and `topic/` are supported. (Default: `SubscriptionSize`, aka `NumUndeliveredMessages`)
+
+- `aggregation` - The aggregation function used to aggregate the metric. (Values: mean, median, variance, stddev, sum, count, percentileX (X: an integer in the range 0 to 100), Required for metrics with value type `DISTRIBUTION`)
 
 
 - `activationValue` - Target value for activating the scaler. Learn more about activation [here](./../concepts/scaling-deployments.md#activating-and-scaling-thresholds).(Default: `0`, Optional, This value can be a float)
@@ -35,6 +40,10 @@ The Google Cloud Platform‎ (GCP) Pub/Sub trigger allows you to scale based on
   - Just the subscription name, in which case you will reference a subscription from the current project or the one specified in the credentials file used.
   - Use the full link provided by Google, so that you can reference a subscription that is hosted in another project Eg: `projects/myproject/subscriptions/mysubscription`.
 
+- `topicName` defines the topic that should be monitored. Similar to `susbcriptionName`, you can use different formulas:
+  - Just the topic name, in which case you will reference a topic from the current project or the one specified in the credentials file used.
+  - Use the full link provided by Google, so that you can reference a topic that is hosted in another project Eg: `projects/myproject/topics/mytopic`.
+
 Here's an [example](https://github.com/kedacore/sample-go-gcppubsub).
 
 ### Authentication Parameters
@@ -65,6 +74,28 @@ spec:
       credentialsFromEnv: GOOGLE_APPLICATION_CREDENTIALS_JSON # Required
 ```
 
+### Example using distribution-valued metric and aggregation
+
+```yaml
+apiVersion: keda.sh/v1alpha1
+kind: ScaledObject
+metadata:
+  name: pubsub-scaledobject
+  namespace: keda-pubsub-test
+spec:
+  scaleTargetRef:
+    name: keda-pubsub-go
+  triggers:
+  - type: gcp-pubsub
+    metadata:
+      # Example: https://cloud.google.com/pubsub/docs/monitoring#monitoring_message_throughput_2
+      mode: "MessageSizes"
+      aggregation: "count" # Calculate count of messages being published
+      value: "50"
+      topicName: "mytopic" # Required
+      credentialsFromEnv: GOOGLE_APPLICATION_CREDENTIALS_JSON # Required
+```
+
 ### Example using TriggerAuthentication
 
 ```yaml