Re:We built network isolation for 1500 services
by marcinzm on Wednesday November 06, 2019 @06:32AM
(#21461522)
I'm curious why the investment in so many services instead of investing in improved CI/CD/build tools. For example, automatically deploying services when the libraries they depend on are updated. Monitoring to automatically check if a deployment is having issues. I mean, due to security concerns you still will need to re-deploy services every so often so having great CI/CD seems a more worthwhile investment.
edit: Also, for the same reason chaos engineering, short lived certificates and periodic backup tests make sense; constantly testing out the deployment of all services makes sense to me. If you do something like deploying a service rarely then the chance of something going wrong (forgot a step, radical external changes, etc.) is decently higher.
edit: Also, for the same reason chaos engineering, short lived certificates and periodic backup tests make sense; constantly testing out the deployment of all services makes sense to me. If you do something like deploying a service rarely then the chance of something going wrong (forgot a step, radical external changes, etc.) is decently higher.