The graphml has everything we need, except IP addresses. We can get those from the backend, but that's only actually needed for scenarios currently.
254 and #278 are already cleaning up IP stuff... ultimately we should be able to get rid of a lot of the metadata we dump in the compose file in environment variables and labels and in the kubernetes deployment several V1EnvVar are set just for the sake of recovering "from deployment".
This also means initiatives like #285 and #288 will be easier to build since we wont need to add extra metadata in the deployments.
The graphml has everything we need, except IP addresses. We can get those from the backend, but that's only actually needed for scenarios currently.
254 and #278 are already cleaning up IP stuff... ultimately we should be able to get rid of a lot of the metadata we dump in the compose file in
environment
variables andlabels
and in the kubernetes deployment severalV1EnvVar
are set just for the sake of recovering "from deployment".This also means initiatives like #285 and #288 will be easier to build since we wont need to add extra metadata in the deployments.