How to scale to 60 million requests (with ZERO ops team) using Cloud Run

Опубликовано 16 апреля 2026, 19:00

Cloud Run → goo.gle/4tMWdaC
RocketFlag → goo.gle/48KnAda
The Batching Strategy →goo.gle/424ZlCP

Martin Omander always says Cloud Run scales well, but does it? In this one, JK joins me to break down how he took a side project (RocketFlag) from zero to *60 million requests per month* without hiring an ops team or losing his mind over a massive cloud bill.

Watch along and learn about:
* *The stack:* Why Cloud Run is a "cheat code" for cold starts and high concurrency workloads.
* *The "cost trap":* How Firestore and BigQuery writes can cost a lot if you aren't careful (and the batching system JK built in an all nighter to fix it).
* *Security:* Using Cloud Armor to block the "garbage traffic" and script kiddies before they even hit your app logs.
* *The bill:* A transparent look at the December invoice; spoiler it’s about $180 USD.

If you’re building something solo and want to see how to handle "real" traffic without becoming a full time Site Reliability Engineer (SRE), this one is for you.

Chapters:
0:00 - Intro
0:25 - Development story
7:36 - Q&A
8:30 - Conclusion

Speakers: Martin Omander, JK Gunnink
Products Mentioned: Cloud Run, Firestore, BigQuery, Cloud Armor, Cloud Billing