AWS DAS passed! The most important cert for Data engineer is behind me. I feel happy and satisfied because I can feel I proved some level of knowledge and I can choose the correct services in the AWS platform at least.
At the end, I shared with you my quick tips if you are already in the middle of DAS cert learning process.
These were busy 3 months of learning, if you’re interested in the following short text I will share my experiences on what the strategy I took, how was the exam and what hints I would give to my past self and for you. It was 3rd certification in the row I took for AWS in the past months, because my company actively supports such initiatives - thanks Nordcloud!
My learning strategy
To make it short, I took some self-paced online trainings, do a handful of example cert exams and did one project. I also tried to be honest with myself and draw honest conclusions from my mistakes.
During the preparation phase, I understood that I will be more motivated when some real project will be made. I am a developer and I love to create. Moreover, lately, I invested in learning how to deal with IoT devices. So I joined 2 worlds by making a simple proof of concepts based on Firehose and prepared a Terraformed infrastructure to load sensors data to the cloud, ingest and transform and display on Athena and Grafana. You can see this project on my GitHub.
Apart from projects I took several MOC in the area of the specialization. Word of caution: don’t rely on one course solely. First, I took A Cloud Guru learning path. That’s one of them in the list of courses here. Unfortunately, after attempting first practice exams I felt I am not prepared at all. I used also courses on Pluralsight and Udemy and have the feeling, the one by Stephane Mareek is the most insightful. Practice exams are an important part of my pre-certification period. There are some included in online courses (in all 3 platforms there were some), however, I found Tutorials Dojo practice exams the most difficult and a few questions from them were exactly the same when I sat the real exam. During the learning from courses, I make detailed notes, but in the last days of learning, I mostly concentrate on the missing points, things that were new in the material I studied. When I did some practice exam, I tried to investigate why I failed, so why I made the wrong answer. If my reasoning path was wrong, I try to understand what led me to this path and what I should learn. On the day of the exam (even 1 day before) I plainly reviewed all my notes to see what I ingested during this time (“ingestion” is a nice ETL-based word :) ). I also review AWS cheat sheets from Tutorials Dojo.
I prefer to attend the exam in the certified Pearson Vue place than doing the proctored exam. Why? Surely, I feel less stress that something will go wrong - any issues regarding power and internet connection are not my responsibility. I also feel at ease, because I can freely talk to myself and look through the window and change focus to make my eyes comfortable. DAS is a pretty long exam (180-220 minutes), so I felt it is a good choice. The only drawback was adjusting to the exam office schedule. Maybe one day I will try the proctored exam. During the exam, I flagged all questions I was not sure of. I was reviewing them again. After that, I had enough time to review all questions and answers. Sometimes, questions were helpful in getting the idea to answer another question. So read carefully and draw conclusions. In some scenario questions, the answers are very similar to each other (f.e. 3 sentences that differ only by one or two services). Usually, out of 4 answers, there were 2 distractors I could eliminate at first glance and then choosing the answer was only 50-50.
I was surprised by the following things during my exam. Disclaimer: it is true to my exam. Yours can be different, e.g. you could have some extensive EMR questions!
- A lot of questions from Lake Formation were present, which is a pretty new AWS offering. When listening to (typically 2-3 yrs old) online courses, it was not there yet.
- EMR was not present almost all, so it was surprising for me, because I prepared for that.
- No questions on CloudWatch metrics, f.e. manual scaling based on metrics and events.
- AWS exam authors are not interested in DynamoDB at all, even in cooperation with analytics services (checkpointing for Kinesis). In general, I felt like the exam was checking the depth of Kinesis and Redshift.
#1 Study connections.
If you know very well how services talk to each other, you are already ahead. Most scenario questions will try to trick you that services can communicate, while they cannot. And if you know connections well, you will immediately scan for the right answer (f.e. when I saw Splunk I was just scanning for Firehose).
#2 Best practices is your friend
Sometimes questions directly refer to best practices. So if you are familiar with them, you can apply them right off the bat. Some examples include the effective use of COPY or partitioning rules. Read and understand best practices for all analytical services, especially all Kinesis family members, Redshift and Glue/Athena.
#3 First eliminate
I discovered that during the exam time the best strategy was first to try to quickly eliminate some distractors and focus on what is left.
#4 Do the practice exams
Not only one, the more, the better. It means you have to spend some money (f.e. 15 USD on tutorials dojo) but I believe this is investing in your future self. Some questions I spotted were repeating through all the practice exams, including my final exam on certification day zero. Some examples are partitioning of Athena files, efficient COPY for Redshift, troubleshooting of Kinesis failures.
I am going to rest for a while and then attempt another certification. Why not. Maybe Database specialty or switch to GCP to learn a different provider. Terraform Associate is also a nice option.
Keep learning and use your knowledge.