systemd: Start request repeated too quickly (start-limit-hit)

What the message looks like

myapp.service: Start request repeated too quickly.
myapp.service: Failed with result 'start-limit-hit'.
Failed to start myapp.service - My application.

Common causes

• The program exits right after start: wrong config, missing file, permission problem or a port that is already in use.
• Restart=always with a very short RestartSec restarts a broken service too often.
• A dependency (database, mount, network) is not ready yet.
• More than 5 starts within 10 seconds, the defaults of StartLimitBurst and StartLimitIntervalSec.
• A bad update changed a binary or a configuration.

Check

1. State and last lines of the unit
   systemctl status myapp --no-pagerCopy
2. Why did it crash? Read the journal of the unit
   journalctl -u myapp -n 50 --no-pagerCopy
3. The limits that apply to this unit
   systemctl show myapp -p Restart -p RestartUSec -p StartLimitBurst -p StartLimitIntervalUSecCopy

Fix

• Find and fix the real error in the journal first (config, permission, port, dependency).
• Clear the failed state: systemctl reset-failed myapp, then systemctl start myapp.
• Give the service breathing room: RestartSec=5 in [Service] and, if needed, StartLimitIntervalSec=60 and StartLimitBurst=5 in [Unit].
• Add After= and Requires= for dependencies that must be ready first.
• Run systemctl daemon-reload after editing the unit file.

FAQ

Can I just raise the start limit?

You can, but it only hides a crash loop. First find out why the process exits. Raise the limit only for a service that is meant to restart often.

How do I see all failed units?

Run systemctl --failed.