Yes, you read it right!
I was able to clear all 3 Google Cloud certifications in 3 weeks. In this post, I will be sharing the resources and my approach to help you prepare for GCP certifications.
A bit about me:
I am a Python+Django developer with hands-on experience of Python-based scripting and Google Cloud Platform working with MediaAgility. I have worked on designing and developing projects to production scale from scratch. At MediaAgility, I am in the Machine Learning team. Since my organization is a Google Cloud Premier Partner, I have also worked on GCP migrations — setting production grade networking, Kubernetes, and GCP infra automation (IAM, networking, Kubernetes, Dataproc, VM, and more).
Google Cloud certifications need a right balance of both theory and practical skills. For the theory, Google Cloud documentation is a great resource and to increase your practical skills, you will have to practice frequently and extensively. You can use free $300 credit from GCP and try different services or join Coursera Google Cloud Architect or Data Engineer course and watch all the videos and complete all the Labs.
Week 1 — Google Cloud Architect:
I started with Coursera’s Architecting with Google Cloud Certification. This exam checks your ability to provide a solution on the Google Cloud. It is almost 80% conceptual and 20% practical. You must know beforehand -
- Various managed services available on Google Cloud
- How different components of Cloud can be fit together to provide a solution
I watched all the videos in just 2–3 days and skipped the concepts that I knew already owing to my experience on Google Cloud. Sharing the important topics below -
- Networking: It is the most important topic and includes everything about Google Cloud Networking. A word of caution, regular practice is needed to get a hang of all the theoretical concepts like This topic is a must. You must be able to write GCloud commands for IAM Roles.
- Shared VPC and VPC peering
- Cloud Router
- Firewall Rules
- Load Balancing
- Storage: You must have a clear understanding of different storage options available in Google Cloud and when to use them. Check this link for more information.
- Compute: This includes App Engine, Compute Engine, and Kubernetes. You need to learn deployment, versioning, and rollback in App Engine. In Kubernetes, you must learn docker files — how to build docker files, different type of docker file and what are the different components of Kubernetes cluster (pod, node, services, load balancing, exposing services to the outside world). You must have a clear understanding of the different components of this construct’s yaml file.
- Deployment: You must gain a good understanding of different pieces of deployment .yaml file. From this section, you can expect some good question related to Deployment manager.
- Monitoring: You should be aware of Stackdriver stack i.e. monitoring, logging, tracing, debugging, profiling. Filtering, logging, and generating monitoring alerts are must-to-know topics.
- Security: This includes working with IAM Roles, Firewall settings, and others. There are some roles or permission which are specific. You must have a clear understanding of the different roles that IAM provides.
Week 2 — Associate Cloud Engineer
This certification tests you for the ability to deploy a solution on Google Cloud. You must have a strong practical experience with Google Cloud — working knowledge of GCloud SDK and Google cloud console.
More details about the exam can be found here.
You must be able to perform the below actions -
- IAM Roles — This topic is a must. You must be able to write GCloud commands for IAM Roles.
- Copying roles from one project to another role
- Custom Roles
- BigQuery IAM roles, AppEngine IAM roles and more
- Billing IAM Roles
- Service Accounts
- Compute/App Engine/Kubernetes: You must learn how to create compute VM from images, how to create disk images, snapshots, how to share images with other projects, and what are the default scopes on Compute VM. Other topics include — deploying app on App Engine, versioning, rollback, the difference between standard and flexible environments. For Kubernetes, you must be able to create docker image, deploy on Kubernetes cluster, create a Kubernetes cluster and write deployment, pod, and services yaml files, and know about mounting storage in Kubernetes.
- Networking: You must learn how to set up VPC, custom mode VPC, and automatic subnet VPC. Also, learn how to create subnets for Compute and Kubernetes (we need to specify pod and services subnet range and secondary subnet range).
Week 3 — Data Engineer
This certification tests your ability to design big data solutions on GCP. This exam expects that you are familiar with the big data products (storage, processing, display) and their open source alternatives as well. This is required because some of the questions expect you to answer GCP alternatives for open source big data products.
BigQuery is the product that you must understand clearly. If you understand BigQuery, you can answer 40% of this exam.
- This course — Data Engineering on Google Cloud Platform Specialization, from Coursera is a good start.
- This article, this article, and this article is good to understand. These are about writing, saving, and sharing queries, moving data in-out from BigQuery.
- Learn about BigQuery data transfer service and use of BigQuery for GeoData.
- Few questions on BigQuery ML.
Sharing an overview of other important topics and a few more resources:
- You must have a clear understanding of Hadoop ecosystem; most of the tools from Hadoop ecosystem have an alternative in GCP like Data Proc for Hadoop Spark cluster, Dataflow for Apache Beam, Composer for Airflow, and others.
- Case Studies:- Refer to the case study section from this link. Make sure you are able to break these case studies and find the GCP alternative.
- You must be able to understand how different GCP components fit with GCP big data products.
- Some questions on Data Studio, viewing BigQuery data in Data Studio along with caching concept.
- Learn about DataFlow input source and sink of Dataflow processed data.
- GCS — You must understand different classes of GCS, moving data from one class to another class, object life cycle, ways to get data into GCS and what components can be used with GCS and use of GCS with DataProc cluster.
- Learn about the use of DataPrep and DataLab in GCP, and target users of these products.
- Machine Learning- This is the area where a data engineer lacks. It is not easy to find a candidate who has knowledge of both Big Data and Machine Learning and that’s what this exam tests you on. You are not expected to be a master of Machine learning, but you must understand below machine learning concepts.
- Bias-Variance trade-off, overfitting and underfitting, training, linear regression, classification, Gradient descent
- Machine learning models, what is the use of GCP MLE (Machine learning engine)
My Schedule -
I studied from 10 p.m. to 2 a.m during those 3 weeks of my preparation. This needed me to take care of my health so that I remain focused. So, I would suggest you maintain a healthy diet and not let the stress get the better of you!
TIPS for Exams and a Few More Resources & Information:
- Make sure you understand the Compute resources on GCP
- You must have a great understanding of Google Cloud Storage
- Stackdriver monitoring is a must. You can expect 3–4 questions easy-to-answer on Stackdriver
- You should clearly understand the IAM roles
- Make sure you understand the GCP architecture and design one give a problem statement. For more information, check this
- You should be able to understand how resources are organized in projects and how projects, folders, and organization structure works in GCP
- This article is a must read
- Make sure you are able to clear GCP practice exam questions
- Don’t worry about the exam duration, the exam is for 2 hours and you will have enough time to review your answers at least one time. If you are not confident about any answer, check the answer you think is the correct one and mark the question as ‘review later’ and come back to the question later with a fresh mind once you finish the other questions. Also, it is possible that you may deduce the correct answer for the previous questions while answering the other questions
If you thoroughly understand the ‘How-to’ and ‘Concepts‘ sections of Google Cloud documentation, you easily have 70% of what it takes to clear GCP certification; remaining 30% is your practice, experience, and your state-of-mind during exams. So, take the exam with a relaxed state of mind.
All The Best!
Here are the links to my certificates.
P.S. Don’t hesitate to click on clap button as many time as you can. :-)