collector races DNS at boot: one psql name-resolution failure for venus.pensanta.com

Single occurrence in whey's 30-day journal, not a pattern, tracked so a second one is recognisable rather than rediscovered. 2026-09-09 23:09:36 at 102s uptime, agent-telemetry-collect.service (user scope): psql: error: could not translate host name "venus.pensanta.com" to address: Temporary failure in name resolution, then subprocess.CalledProcessError exit status 2 from collect.py:345. The collector timer fired before the resolver was up. Fix if it recurs: order the unit after network-online.target / nss-lookup.target, or give apply() a bounded retry on resolution failure. NOTE FOR ANYONE SEARCHING journald: this is a USER-scope unit, so journalctl -u agent-telemetry-collect.service returns No entries - the records sit under _SYSTEMD_UNIT=user@1000.service with _SYSTEMD_USER_UNIT=agent-telemetry-collect.service, readable via journalctl -o json.