There's nothing wrong in the logs. It just starts as normal, but I think somehow systemd it's thinking it's timing out.
2020/07/28 06:14:21 DEBUG : rclone: Version "v1.52.2-133-g9058ec32-beta" starting with parameters ["/usr/bin/rclone" "serve" "http" "disk:/backup" "--config=/root/.config/rclone/rclone.conf" "--addr" ":80" "--no-modtime" "--user" "xxxxx" "--pass" "xxxx" "--vfs-read-chunk-size" "10M" "--vfs-read-chunk-size-limit" "0" "--timeout=11m" "--contimeout" "3m" "--log-level" "DEBUG" "--log-file" "/opt/rclone.log"]
2020/07/28 06:14:21 DEBUG : Using config file from "/root/.config/rclone/rclone.conf"
2020/07/28 06:14:21 INFO : Using --user random42jand01 --pass XXXX as authenticated user
2020/07/28 06:14:21 INFO : Local file system at /backup: poll-interval is not supported by this remote
2020/07/28 06:14:21 DEBUG : Adding path "vfs/forget" to remote control registry
2020/07/28 06:14:21 DEBUG : Adding path "vfs/refresh" to remote control registry
2020/07/28 06:14:21 DEBUG : Adding path "vfs/poll-interval" to remote control registry
2020/07/28 06:14:21 NOTICE: Local file system at /backup: Serving on http://[::]:80/
It's more of a systemd question with rclone, than a issue with rclone I think
But since I don't know the behavior of rclone after it completes, I don't know which one it would be appropriate. I just went with notify because it's what I use for the mount
Ah, you don't want notify - rclone has special support for that in the mount which isn't in the serve command. Probably simple is the right one I think. I think that is the default so you can leave the line out completely.