Default Site receives incorrect value when multiple SNMP discovery policies are defined

[Christophoclese]

When multiple SNMP discovery policies are defined, each with their own default Site, it appears discovered resources will all be created under the Site defined by the last policy to execute.

Orb Agent Configuration:

orb:
  config_manager:
    active: local
  backends:
    snmp_discovery:
    common:
      diode:
        target: grpc://dev.netbox.internal.domain.tld:8080/diode
        client_id: ${DIODE_CLIENT_ID}
        client_secret: ${DIODE_CLIENT_SECRET}
        agent_name: any name
  policies:
    snmp_discovery:

      snmp_tuk_core:
        config:
          retries: 3
          defaults:
            site: Tukwila Lab
            role: Core Switch
          lookup_extensions_dir: /opt/orb/snmp-extensions
        scope:
          targets:
          - host: "10.16.6.1"
          authentication:
            protocol_version: "SNMPv2c"
            community: "community string"

      snmp_ptl_core:
        config:
          retries: 3
          defaults:
            site: Portland
            role: Core Switch
          lookup_extensions_dir: /opt/orb/snmp-extensions
        scope:
          targets:
          - host: "10.17.6.1"
          authentication:
            protocol_version: "SNMPv2c"
            community: "community string"

      snmp_boi_core:
        config:
          retries: 3
          defaults:
            site: Boise
            role: Core Switch
          lookup_extensions_dir: /opt/orb/snmp-extensions
        scope:
          targets:
          - host: "10.18.6.1"
          authentication:
            protocol_version: "SNMPv2c"
            community: "community string"

[Christophoclese]

Devices actually aren't just being assigned the Site of the last policy to execute, but rather every site defined in policy is applied to all of the discovered resources one Site at a time. I was able to refresh NetBox and watch the Site value for the devices slowly update to the various Site names. There appear to be no Changelog entries for these actions, however.

[Christophoclese]

Adding a Location to the defaults seems to workaround this issue, although while Diode was ingesting the updates from Orb, when refreshing the view of devices, I still noticed the site on some devices temporarily displaying an incorrect value. They eventually settled on the expected site values.

[...]
      snmp_tuk_core:
        config:
          retries: 3
          defaults:
            site: Tukwila Lab
            role: Core Switch
            location: MDF
          lookup_extensions_dir: /opt/orb/snmp-extensions
        scope:
          targets:
          - host: "10.16.6.1"
          authentication:
            protocol_version: "SNMPv2c"
            community: "community string"

      snmp_ptl_core:
        config:
          retries: 3
          defaults:
            site: Portland
            role: Core Switch
            location: MDF
          lookup_extensions_dir: /opt/orb/snmp-extensions
        scope:
          targets:
          - host: "10.17.6.1"
          authentication:
            protocol_version: "SNMPv2c"
            community: "community string"

      snmp_boi_core:
        config:
          retries: 3
          defaults:
            site: Boise
            role: Core Switch
            location: MDF
          lookup_extensions_dir: /opt/orb/snmp-extensions
        scope:
          targets:
          - host: "10.18.6.1"
          authentication:
            protocol_version: "SNMPv2c"
            community: "community string"
[...]