diff --git a/deploy/fleetd.service b/deploy/fleetd.service
index 38e42ef..92f2de5 100644
--- a/deploy/fleetd.service
+++ b/deploy/fleetd.service
@@ -12,8 +12,16 @@
 #   # edit WorkingDirectory / ExecStart below for your host's paths and java location
 #   systemctl --user daemon-reload
 #   systemctl --user enable --now herdr fleetd
+#   loginctl enable-linger $USER      # REQUIRED -- see below
 #   journalctl --user -u fleetd -f
 #
+# `loginctl enable-linger` is not optional and is easy to miss, because leaving it out looks like
+# success: `systemctl --user enable` reports "enabled" and both units run for as long as you stay
+# logged in. A user manager without lingering starts at your first login and stops at your last
+# logout, so the fleet simply does not come back after a reboot -- which is the whole reason to
+# use systemd here rather than the setsid scripts these units replaced. Check it with
+# `loginctl show-user $USER -p Linger`; the answer must be `Linger=yes`.
+#
 # Secrets (AI_GATEWAY_TOKEN, WORKER_GITEA_TOKEN, LAVINMQ_URI, COORD_AMQP_URI, ...) are not set
 # here and need no systemd drop-in: ExecStart runs a login shell, so they come from wherever your
 # login shell already sources them (this host: ~/.fleet/secrets.sh via ~/.zprofile). If a token is