Alerting policies in GCP - Commands & Configuration

Introduction

Alerting policies help you get notified when something important happens in your cloud resources. They watch metrics or logs and send alerts so you can fix problems quickly.

When you want to know if your website is down or slow

When your server CPU usage goes too high for too long

When a storage bucket is almost full

When an error rate in your application increases suddenly

When you want to get notified by email or SMS about critical issues

Config File - alert_policy.yaml

alert_policy.yaml

displayName: "High CPU Usage Alert"
combiner: "OR"
conditions:
- displayName: "CPU usage above 80%"
  conditionThreshold:
    filter: "metric.type=\"compute.googleapis.com/instance/cpu/utilization\""
    comparison: COMPARISON_GT
    thresholdValue: 0.8
    duration: "300s"
    trigger:
      count: 1
notificationChannels:
- projects/my-project/notificationChannels/1234567890
enabled: true

displayName: The name of the alert policy.

combiner: How multiple conditions combine (OR means alert if any condition is true).

conditions: What metric to watch and the threshold to trigger alert.

notificationChannels: Where to send alerts (email, SMS, etc.).

enabled: Whether the alert is active.

Commands

This command creates an alerting policy in Google Cloud Monitoring using the configuration file. It sets up the alert to watch CPU usage and notify when it goes above 80%.

Terminal

gcloud monitoring policies create --policy-from-file=alert_policy.yaml

Expected Output

Created alerting policy: projects/my-project/alertPolicies/9876543210

→

--policy-from-file - Specifies the YAML file that defines the alerting policy

This command lists all alerting policies in your project so you can verify the new policy was created.

Terminal

gcloud monitoring policies list

Expected Output

NAME                  DISPLAY NAME           ENABLED

9876543210            High CPU Usage Alert   True

This command shows detailed information about the alerting policy to check its settings and conditions.

Terminal

gcloud monitoring policies describe 9876543210

Expected Output

name: projects/my-project/alertPolicies/9876543210

displayName: High CPU Usage Alert

combiner: OR

conditions:

- displayName: CPU usage above 80%

  conditionThreshold:

    filter: metric.type="compute.googleapis.com/instance/cpu/utilization"

    comparison: COMPARISON_GT

    thresholdValue: 0.8

    duration: 300s

    trigger:

      count: 1

notificationChannels:

- projects/my-project/notificationChannels/1234567890

enabled: true

Key Concept

If you remember nothing else from this pattern, remember: alerting policies watch your cloud metrics and notify you automatically when something needs your attention.

Common Mistakes

Not enabling the alerting policy after creating it

The alert will not send notifications if the policy is disabled.

Make sure the 'enabled' field is set to true in the policy configuration.

Using incorrect metric filter syntax in the condition

The alert will never trigger because it is not watching the right metric.

Use the exact metric type string from Google Cloud Monitoring documentation, like "compute.googleapis.com/instance/cpu/utilization".

Not specifying notification channels

Alerts will trigger but no one will be notified without notification channels.

Create notification channels in Google Cloud Console and include their IDs in the policy.

Summary

Create an alerting policy YAML file defining the metric, threshold, and notifications.

Use gcloud CLI to create the alerting policy from the YAML file.

List and describe policies to verify the alert is set up correctly and enabled.

Practice

1. What is the main purpose of an alerting policy in Google Cloud Monitoring?

easy

A. To watch cloud metrics and notify when something needs attention

B. To deploy virtual machines automatically

C. To store data backups securely

D. To manage user access permissions

Solution

1. Step 1: Understand the role of alerting policies

   An alerting policy monitors cloud metrics continuously to detect issues.

2. Step 2: Identify the main function

   When a condition is met, it sends notifications to inform the team.

3. Final Answer:

   To watch cloud metrics and notify when something needs attention -> Option A

4. Quick Check:

   Alerting policy = notify on metric issues [OK]

Hint: Alerting policies notify on metric changes [OK]

Common Mistakes:

• Confusing alerting with deployment
• Thinking alerting manages backups
• Mixing alerting with access control

2. Which of the following is the correct way to specify a notification channel in a Google Cloud alerting policy configuration?

easy

A. "channels": "my-project/notification/12345"

B. "notifyChannel": "12345"

C. "notification": "channel-12345"

D. "notificationChannels": ["projects/my-project/notificationChannels/12345"]

Solution

1. Step 1: Recall notification channel format

   Google Cloud expects notification channels as full resource names in an array.

2. Step 2: Match correct syntax

   "notificationChannels": ["projects/my-project/notificationChannels/12345"] uses the correct JSON key and full resource path format.

3. Final Answer:

   "notificationChannels": ["projects/my-project/notificationChannels/12345"] -> Option D

4. Quick Check:

   Notification channels use full resource names in array [OK]

Hint: Use full resource names in an array for notificationChannels [OK]

Common Mistakes:

• Using wrong key names like notifyChannel
• Providing notification channel as a string, not array
• Omitting project prefix in resource name

3. Given this alerting policy condition snippet:

{"conditionThreshold": {"filter": "metric.type=\"compute.googleapis.com/instance/cpu/utilization\"", "comparison": "COMPARISON_GT", "thresholdValue": 0.8, "duration": "300s"}}

What triggers this alert?

medium

A. CPU utilization is less than 80% for 5 minutes

B. CPU utilization is greater than 80% for 5 minutes

C. CPU utilization is exactly 80% at any time

D. CPU utilization is greater than 80% for 30 seconds

Solution

1. Step 1: Interpret the condition filter and comparison

   The filter targets CPU utilization metric, and comparison is greater than 0.8 (80%).

2. Step 2: Understand the duration

   The duration is 300 seconds, meaning the condition must hold for 5 minutes continuously.

3. Final Answer:

   CPU utilization is greater than 80% for 5 minutes -> Option B

4. Quick Check:

   Threshold > 0.8 for 300s = alert [OK]

Hint: Check comparison and duration for alert trigger [OK]

Common Mistakes:

• Confusing greater than with less than
• Ignoring duration requirement
• Assuming exact equality triggers alert

4. You created an alerting policy but no notifications are sent when the condition is met. What is the most likely cause?

medium

A. Notification channels are not configured or linked properly

B. The metric filter is too broad

C. The alerting policy is disabled by default

D. The project has no billing account

Solution

1. Step 1: Check notification channel setup

   If no notifications arrive, the most common cause is missing or incorrect notification channels.

2. Step 2: Verify alerting policy status and billing

   Alerting policies are enabled by default and billing issues usually don't block notifications.

3. Final Answer:

   Notification channels are not configured or linked properly -> Option A

4. Quick Check:

   No notifications usually mean missing channels [OK]

Hint: Verify notification channels linked to policy [OK]

Common Mistakes:

• Assuming alerting policies are disabled by default
• Blaming billing for notification failure
• Ignoring notification channel configuration

5. You want to create an alerting policy that notifies your team only if CPU usage is above 70% for 10 minutes, but only during business hours (9 AM to 5 PM). Which approach best achieves this?

hard

A. Set the alert threshold to 70% and rely on team to ignore alerts outside business hours

B. Create an alerting policy without conditions and manually check CPU usage during business hours

C. Use an alerting policy with a condition on CPU usage and add a time window using an uptime check or schedule-based notification channel

D. Configure the alerting policy to notify continuously regardless of time

Solution

1. Step 1: Define CPU usage condition with duration

   Set condition threshold to >70% CPU usage sustained for 10 minutes.

2. Step 2: Limit notifications to business hours

   Use scheduling features like uptime checks or notification channel schedules to restrict alerts to 9 AM-5 PM.

3. Final Answer:

   Use an alerting policy with a condition on CPU usage and add a time window using an uptime check or schedule-based notification channel -> Option C

4. Quick Check:

   Combine metric condition + time window for precise alerts [OK]

Hint: Combine metric condition with time-based notification [OK]

Common Mistakes:

• Ignoring time restrictions in alerting policy
• Relying on manual checks instead of automation
• Not setting duration for CPU usage condition