🚀 Journey to the Golden Kubestronaut (6/15): PCA Passed, Experiencing the New World of Monitoring! (feat. The Struggle with PromQL)

Hello everyone! This is gasbugs, taking one step at a time towards the grand goal of becoming a ‘Golden Kubestronaut’. Not long after achieving the first milestone of ‘Kubestronaut’ by passing CKS, I immediately challenged the sixth hurdle, the PCA (Prometheus Certified Associate) exam, and I’m thrilled to share the news of my success! 🥳

Having mastered how to secure a cluster (CKS), I thought it was time to gain the ‘eyes’ to thoroughly inspect and observe the cluster’s status. At the core of this, undoubtedly, was Prometheus. But to be honest, this challenge was nothing short of ‘reckless’ itself.

🤯 “I’m going to fail this…” A One-Day Challenge, and the Fateful 86 Points

To be honest, I haven’t directly handled practical tasks like Prometheus rule setting recently. However, with a vague confidence that ‘I’ve done a lot of it in the past, so it’ll be fine,’ I dedicated just one day before the exam to cramming. As a result, I barely passed with 86 points against a cutoff of 75, a score where luck played a bigger role than skill. Throughout the exam, I thought, ‘Ah, this is a disaster,’ so I can’t express how surprised I was to see the pass screen.

This time, too, I quickly reviewed the concepts with the help of my reliable AI pair programmer, Gemini. However, PCA was definitely not an exam that could be passed with concepts alone. Looking back, I probably would have passed stably if I had gone through the process of solving at least 200-300 practical problems and becoming one with PromQL. I strongly advise against reckless challenges like mine!

(It seems that PromQL-related questions accounted for roughly 20 out of the total 60 questions.)

🎯 PCA Specifically Asked This: Memorable Killer Questions

The PCA exam persistently delved into how deeply one understands Prometheus’s philosophy and core functionalities. The key questions I encountered in the exam were as follows:

1. PromQL: Grasp the ‘Intent’ beyond the Syntax: It wasn’t just about asking for simple function usage.

• Do you accurately know when and why to use functions like avg_over_time, sum_over_time? – In particular, many questions were asked about understanding the difference in meaning of results based on query order, such as “to prevent statistical errors, summation (sum) must be performed before calculating ratios.” A deep understanding of the data model was essential.
• Alerting Rule: Analyze ‘Where’ it’s Heading: The exam presented complex alerting_rules.yml files and required the ability to accurately trace the flow of how a specific alert is triggered under certain conditions and ultimately which receiver it matches with to send notifications. A thorough understanding of Alertmanager’s routing mechanism was crucial.

3. Include ‘Dynamic’ Information in Custom Alerts: Beyond simply triggering an alert, the exam asked if you knew how to dynamically fetch and include label variables from the triggered metric in the alert message (Annotation) when a custom rule is fired. The ability to use template syntax like {{ $labels.instance }} was fundamental.
4. Correlation between Metrics and Rules: In addition to these, I also remember questions about basic architecture and operational capabilities, such as whether you can accurately count the number of metric types collected in specific situations, and how many Alert Rules a Prometheus server can connect to (evaluate).

✨ Overall Review: An Exam to Understand the Heart of Observability

PCA was not merely an exam asking if you knew Prometheus’s functionalities. It was a dense exam that evaluated whether you understood the core philosophy of an Observability system and could apply it in practice, asking questions like ‘Why should these metrics be collected?’, ‘How should this data be processed to gain meaningful insights?’, and ‘How can accurate alerts be generated in a timely manner?’

If CKS gave me the ‘shield’ to protect the cluster, then PCA feels like I’ve gained the ‘eyes’ to see through it. I have now passed 6 out of 15 hurdles on the journey to becoming a ‘Golden Kubestronaut’. There’s still a long way to go, but this process of growing step by step and evolving into an expert in the cloud-native ecosystem is truly enjoyable.

Please look forward to and support my seventh journey as well! May you all continue your enjoyable cloud-native voyage! 🎉

---

Tags: PCA, PrometheusCertifiedAssociate, Prometheus, PromQL, Monitoring, Certification Review, CNCF, GoldenKubestronaut, Kubestronaut, Observability