Open nascentcore-eng opened 1 year ago
we can use an easier way, like add an cluster
command:
# add our dev cluster
starship-cli cluster add --name starship-dev --api-address http://my-starship-api-server-0.com:8080
# add our test cluster
starship-cli cluster add --name starship-test --api-address http://my-starship-api-server-1.com:8080
# than, wen can switch out cluster using `use` sub-command
starship-cli cluster use starship-dev
# finally, we can delete cluster:
starship-cli cluster delete starship-dev
# also, we need to have `list`
starship-cli cluster list
In this way, It can reduce our handling of kubeconfig
and also reduce some of the workload for us.
Is your feature request related to a problem? Please describe. Right now, users must specify API Server's address in --api-address flag. Typically, this requires using kubectl to get the
api-server
service, or usingkubectl port-forward <api-server-pod>
And then copy the address to the command line.Describe the solution you'd like CLI reads kubectl's kubeconfig, and automatically figure out how to connect with the installed Starship's API Server.
To allow this, we should also change helm-charts to create a fixed namespace
tricorder
. (It seems not possible https://stackoverflow.com/questions/51783651/how-to-create-a-namespace-if-it-doesnt-exists-from-helm-templates, so we should consider just using yaml templates)Describe alternatives you've considered There seems no obvious alternatives
Additional context There seems no meaningful additional contexts.