Health checks, not hope
A process is restarted because a check failed, not because it happened to exit; groups start and stop together.
Process supervision
supervice
Keep a fleet of processes running, without a process manager to install.
Available
What it is
An async supervisor for the processes an application needs beside it: N workers, daemons, queue consumers and sidecars, with health checks, process groups and restart-on-crash. It has no dependencies of its own, so it adds a supervisor without adding an operational surface. Plain systemd or Kubernetes remain the right answer when they genuinely fit better.
Who it is for. Python services that need to run and supervise several long-lived processes in one place, and would rather not adopt supervisord or a Procfile runner to do it.
- Python
- Supervisor
- Workers
- Health checks
- Restart-on-crash
What it does
No dependencies
Nothing to install beside it, so supervising a fleet does not enlarge the surface you have to operate.
Honest about its limits
Where systemd or Kubernetes is the better fit, the documentation says so rather than arguing for itself.
Ready to look at supervice?
For the documentation, a pilot or an integration, talk to us directly.