artemis: fetch event log before error analysis during provisioning

When a guest enters error state during _check_ip_ready, the event_log_error property attempted to read the event log file from disk. However, dump_events — which fetches events from the Artemis API and writes them to disk — only ran on the success path or during destroy(). The file never existed at this point, so event log analysis was always skipped with a warning, and the error message fell back to the generic "Artemis resource ended in 'error' state".

Fix by calling dump_events before accessing event_log_error in _check_ip_ready, ensuring the event log is available for analysis.

Assisted-by: Claude Code Signed-off-by: Miroslav Vadkerti mvadkert@redhat.com