I’m currently evaluating whether Nomad is a good fit for my project and have a couple questions after finishing the first couple chapters of the docs.
Are jobs persistent in the sense that a job is still known upon death/restart of a nomad server
Does Nomad detect that a server went down and automatically re-deploys the jobs running on the dead server to another one?
Can jobs be submitted via API?
Is there an artificial limit of jobs? My use case would be running N jobs (hundreds or thousands) with basically the same spec (image, resource requirements etc) but varying configuration on a growable cluster
Can Nomad servers seamlessly get added to increase cluster capacity?
Jobs are deployed to clients. Servers manage clients. As long as there is a client available that meets the requirements of the job, and has capacity, yes, the job should get deployed to a new client.
Yes!
No artificial limit. Just capacity limits based on your available resources.
Yes, but I think what you are really looking for is Clients more than servers. You might need to increase the server count eventually, from 3->5-7. But people run hundreds to thousands of clients with 3-to-5 servers routinely.