Open stradig opened 4 months ago
This is a followup to https://www.notion.so/neondatabase/INC-151-Stuck-neon-VMs-and-controller-panics-in-ap-southeast-1-facdaac1ebf2459b8f969f6608f36d40?d=fbb602d92cc44f709ccb1b2389172f4c#5e8c974431f247f583a0f687f8c7fdfb
It was proposed to kill a VM during testing in staging on purpose to check how the scheduling copes with that. We need to investigate the feasibility and how sense it makes.
This will be caught in CI once https://github.com/neondatabase/autoscaling/issues/580 is implemented.
Problem description / Motivation
This is a followup to https://www.notion.so/neondatabase/INC-151-Stuck-neon-VMs-and-controller-panics-in-ap-southeast-1-facdaac1ebf2459b8f969f6608f36d40?d=fbb602d92cc44f709ccb1b2389172f4c#5e8c974431f247f583a0f687f8c7fdfb
It was proposed to kill a VM during testing in staging on purpose to check how the scheduling copes with that. We need to investigate the feasibility and how sense it makes.