Operational excellence in GCP - Commands & Configuration

Introduction

Operational excellence helps you run your cloud systems smoothly and safely. It focuses on making sure your services work well, fix problems quickly, and improve over time.

When you want to monitor your cloud applications to catch issues early.

When you need to automate responses to common problems to reduce downtime.

When you want to keep your cloud resources organized and secure.

When you want to learn from incidents and improve your system's reliability.

When you want to set up alerts to notify your team about important events.

Commands

This command creates a logging sink that exports logs from your App Engine application to a Cloud Storage bucket. It helps you keep logs for analysis and troubleshooting.

Terminal

gcloud logging sinks create my-sink storage.googleapis.com/my-bucket --log-filter="resource.type=gae_app"

Expected Output

Created sink [my-sink].

→

--log-filter - Filters logs to only include those from App Engine applications.

This command creates an alerting policy that notifies you when a Compute Engine instance's CPU usage goes above 80% for more than 60 seconds. It helps you react quickly to performance issues.

Terminal

gcloud monitoring policies create --notification-channels=projects/my-project/notificationChannels/1234567890 --condition-display-name="High CPU Usage" --condition-filter="metric.type=compute.googleapis.com/instance/cpu/utilization AND resource.type=gce_instance" --condition-threshold-value=0.8 --condition-comparison=COMPARISON_GT --condition-duration=60s

Expected Output

Created alerting policy [projects/my-project/alertPolicies/9876543210].

→

--notification-channels - Specifies where alerts are sent, like email or SMS.

→

--condition-filter - Defines which metric and resource to monitor.

This command deploys infrastructure using a Deployment Manager configuration file. It helps you manage resources consistently and track changes.

Terminal

gcloud deployment-manager deployments create my-deployment --config deployment.yaml

Expected Output

Create operation operation-1234567890-abcdef started.

NAME          TYPE                     STATE      ERRORS

my-deployment deploymentmanager.v2beta.type RUNNING    []

→

--config - Specifies the configuration file describing the resources.

This command shows details about your deployment, including status and any errors. It helps you verify that your infrastructure is set up correctly.

Terminal

gcloud deployment-manager deployments describe my-deployment

Expected Output

name: my-deployment

operation:

  status: DONE

  error: null

  operationType: insert

  targetLink: https://www.googleapis.com/deploymentmanager/v2/projects/my-project/global/deployments/my-deployment

  user: [e-mail]

Key Concept

If you remember nothing else from this pattern, remember: monitoring, alerting, and managing resources help keep your cloud systems reliable and easy to fix.

Common Mistakes

Not setting up log sinks or alerting policies after deploying applications.

Without logs and alerts, you won't know when problems happen or how to fix them quickly.

Always create log sinks and alerting policies to monitor your applications and get notified of issues.

Using overly broad log filters that collect too much data.

This can cause high storage costs and make it hard to find important information.

Use specific filters to collect only the logs you need for troubleshooting.

Ignoring deployment status and errors after running deployment commands.

You might miss failed resource creations or misconfigurations that cause downtime.

Always check deployment status and fix any errors reported.

Summary

Create log sinks to export and keep important logs for troubleshooting.

Set up alerting policies to get notified about critical issues like high CPU usage.

Use deployment tools to manage cloud resources consistently and check deployment status.

Practice

1. What is the main goal of operational excellence in cloud systems?

easy

A. Keep systems healthy and reliable

B. Write complex code

C. Reduce cloud costs only

D. Avoid using monitoring tools

Solution

1. Step 1: Understand operational excellence purpose

   Operational excellence focuses on keeping cloud systems running well and reliable.

2. Step 2: Compare options with this goal

   Only Keep systems healthy and reliable matches this goal; others focus on unrelated tasks.

3. Final Answer:

   Keep systems healthy and reliable -> Option A

4. Quick Check:

   Operational excellence = system health and reliability [OK]

Hint: Operational excellence means system health and reliability [OK]

Common Mistakes:

• Confusing cost reduction with operational excellence
• Thinking it means writing complex code
• Ignoring monitoring and alerts

2. Which GCP service is used to monitor cloud resources and send alerts?

easy

A. Cloud Monitoring

B. Cloud Storage

C. Cloud Functions

D. Cloud SQL

Solution

1. Step 1: Identify monitoring service in GCP

   Cloud Monitoring is designed to track resource health and send alerts.

2. Step 2: Eliminate unrelated services

   Cloud Storage stores data, Cloud Functions run code, Cloud SQL manages databases.

3. Final Answer:

   Cloud Monitoring -> Option A

4. Quick Check:

   Monitoring and alerts = Cloud Monitoring [OK]

Hint: Monitoring and alerts in GCP use Cloud Monitoring [OK]

Common Mistakes:

• Choosing Cloud Storage for monitoring
• Confusing Cloud Functions with monitoring
• Selecting Cloud SQL instead of monitoring

3. Given this snippet of a GCP alert policy configuration:

{
  "conditions": [{
    "displayName": "High CPU Usage",
    "conditionThreshold": {
      "filter": "metric.type=\"compute.googleapis.com/instance/cpu/utilization\"",
      "comparison": "COMPARISON_GT",
      "thresholdValue": 0.8,
      "duration": "300s"
    }
  }]
}

What does this alert do?

medium

A. Triggers if CPU usage is above 80% for 30 seconds

B. Triggers if CPU usage is below 80% for 5 minutes

C. Triggers if CPU usage is below 80% for 30 seconds

D. Triggers if CPU usage is above 80% for 5 minutes

Solution

1. Step 1: Analyze threshold and duration

   ThresholdValue 0.8 means 80%, duration 300s means 5 minutes.

2. Step 2: Understand comparison operator

   COMPARISON_GT means greater than, so alert triggers if CPU > 80% for 5 minutes.

3. Final Answer:

   Triggers if CPU usage is above 80% for 5 minutes -> Option D

4. Quick Check:

   CPU > 80% for 300s = alert triggers [OK]

Hint: COMPARISON_GT with 0.8 threshold means CPU > 80% [OK]

Common Mistakes:

• Misreading COMPARISON_GT as less than
• Confusing duration seconds with minutes
• Ignoring threshold value meaning

4. You wrote this alert policy JSON but it never triggers:

{
  "conditions": [{
    "displayName": "Low Disk Space",
    "conditionThreshold": {
      "filter": "metric.type=\"compute.googleapis.com/instance/disk/bytes_used\"",
      "comparison": "COMPARISON_LT",
      "thresholdValue": 1000000000,
      "duration": "600s"
    }
  }]
}

What is the likely problem?

medium

A. Threshold value is too high

B. Using COMPARISON_LT instead of COMPARISON_GT

C. Filter metric type is incorrect

D. Duration is too short

Solution

1. Step 1: Understand alert intent

   Low disk space alert should trigger when disk usage is high, so comparison should be greater than.

2. Step 2: Check comparison operator

   COMPARISON_LT triggers when usage is less than threshold, opposite of intent.

3. Final Answer:

   Using COMPARISON_LT instead of COMPARISON_GT -> Option B

4. Quick Check:

   Low disk space needs COMPARISON_GT [OK]

Hint: Low disk space alert needs COMPARISON_GT, not LT [OK]

Common Mistakes:

• Using less than instead of greater than
• Misunderstanding metric meaning
• Ignoring alert duration importance

5. Your team wants to improve operational excellence by automating recovery when a VM instance becomes unresponsive. Which GCP service combination best supports this goal?

hard

A. Cloud Storage + Cloud SQL for backups

B. Cloud Pub/Sub + BigQuery for data analysis

C. Cloud Monitoring + Cloud Functions to restart VM automatically

D. Cloud Run + App Engine for hosting apps

Solution

1. Step 1: Identify automation for recovery

   Cloud Monitoring detects issues; Cloud Functions can run code to fix problems automatically.

2. Step 2: Evaluate other options

   Cloud Storage and SQL are for data, Pub/Sub and BigQuery for messaging and analysis, Cloud Run and App Engine for app hosting, not recovery automation.

3. Final Answer:

   Cloud Monitoring + Cloud Functions to restart VM automatically -> Option C

4. Quick Check:

   Monitoring + automation = Cloud Monitoring + Cloud Functions [OK]

Hint: Use Monitoring to detect and Functions to automate fixes [OK]

Common Mistakes:

• Choosing storage or database services for automation
• Confusing data analysis tools with recovery tools
• Ignoring automation in operational excellence