A systemd burst limit configured in 2023 as a precaution, which took three years to be needed.
[Unit]
StartLimitIntervalSec=300
StartLimitBurst=5
[Service]
Restart=always
RestartSec=5
# five restarts in forty seconds, then failed, then an
# alert. as designed, three years later.
Without the limit the worker would have restarted every five seconds indefinitely while reporting active, which is the failure this converts into something visible. A configuration that does nothing for three years and then does exactly the right thing is the argument for defaults chosen carefully rather than reactively.