What's the worst JVM outage you've ever debugged at 3 AM?

Launching JVM CodeLens soon — an AI-powered JVM diagnostic tool that reads your heap dumps, GC logs, and thread dumps and tells you which line of code is crashing production.

Before I do, I want to learn from this community: what's the nastiest Java production incident you've debugged?

Bonus points for:

- How long did it take to find the root cause?

- Which tools helped (or failed you)?

- What would have made the debugging 10x faster?

I'll share the top 3 stories in my launch post with credit. Would love this to become a knowledge-base thread for every Java engineer who inherits an on-call shift.