Chirpstack 3.* provisioning/de-provisioning issues

Echoing @KingCrab comments on #94. I've encountered the same problems running against Lora App Server 3.8.0.

1. This seems to be caused by an old Gateway registration option on an older version of Chirpstack. Documentation doesn't seem to be updated, but I can't find the option anywhere https://www.chirpstack.io/network-server/features/gateway-management/ .

Gateways can be created either by enabling create on stats (see Gateway Configuration) or by using the API.

2. Looks like there was a data race on the deprovision step that was causing almost everything to remain.

Original comments by @KingCrab

I've tried lorhammer with the updated API parts with loraserver 3.1.0 and see a few strange behaviours. Not sure if these are bugs or wrong usage:

1. No gateways are added to the server, whereas everything else (device profiles, devices) was. As a result no data was receided by the server and the orchestrator terminates with "ERRO[0022] Local output error when wait error="signal: interrupt" logger=orchestrator/deploy/local"
2. Devices and Device profiles never get removed from server. Resulting in ~20 LorhammerDeviceProfiles, all with the same name.