[encoderer]

Not the OP but I have some… similar experience. When you run a high availability service without a full ops team, reliable infrastructure is non-negotiable. Burn out has to be managed.