0xABAB

@rpodgorny +1 on following risk management 101.

@victorcoder I meant a particular mode of the system is which no new jobs will be executed. E.g. suitable for when one wants to upgrade dkron.

No, semantically that is different. I am talking about an API call to tell the cluster to stop accepting any new jobs, but finish all the currently running ones. If...

@kevana If you set all existing jobs to disabled, people can still define new jobs and for those jobs nothing would be disabled.

@kevana I am very much interested in something which can realistically consistently get 99.99% availability. What's your relation to this project and perhaps you can comment on #242? It can...