[Bug]: A missing inventory group for the Beats output fails with an unreadable error

beats_filebeat_output defaults to logstash, and the templates resolve the output hosts straight from the inventory group. Without that group the run dies while rendering the config:

TASK [beats : Configure Filebeat]
AnsibleUndefinedVariable: 'dict object' has no attribute 'logstash'

The message names neither the variable nor the group, and it arrives after the packages are installed and the certificates are distributed. #102 was this error and took a maintainer to explain.

Not about the default: elasticsearch behaves the same way. Nine group lookups across the three templates are unguarded, while the Kibana reference in the same files already checks | default([]) | length > 0 and skips its block.

Want

A preflight assert in roles/beats/tasks/main.yml, before the three import_tasks, for every enabled beat whose output is logstash or elasticsearch, saying which variable is set to what and which group is missing.

Only under elasticstack_full_stack, standalone resolves hosts from beats_target_hosts. No fallback: logstash and kibana fall back to localhost because those services sit next to Elasticsearch, a Beat doing that would ship data nowhere.

Read the output as beats_<beat>_output.type | default(beats_<beat>_output) so it keeps working after #522 turns the variable into a dict, and skip types that resolve no group.

Start with roles/beats/tasks/main.yml and the three Beats templates; trace the nine output-host group lookups and the existing Kibana guard. Exercise a run with a missing logstash or elasticsearch inventory group under elasticstack_full_stack. Done when preflight stops before the imports with an actionable variable and group message, while valid groups continue normally.