Hacker News new | ask | show | jobs
by peterwwillis 2700 days ago
It's pretty common practice for complex systems. If it's already dead, and nobody knows what's wrong or how to fix it, you try restarting it. (normally you first redirect traffic to a different region, but apparently big orgs are still running critical infrastructure with changes that affect all regions and can't be backed out)
1 comments

Kubernetes doesn't even have a "restart this deployment" ability...