About
FleetOps is built by Pravin Banjade for the people who keep shared Docker hosts and clusters running: ops and DevOps engineers on small and mid-size teams, working with a terminal in one window and an incident channel in the other.
Why it exists
Container UIs tend to be slow to tell you what changed. You refresh, you poll, you open a terminal anyway. FleetOps starts from the opposite end: the interface should be the fastest way to know what is happening on a host, and the safest way to change it with a team watching.
So it is one Rust binary with the UI and database built in, a log viewer that stays fast at hundreds of thousands of lines, a terminal that never quietly starts another shell, and an audit trail for every change.
What it holds to
- Live by default
- The screen shows Docker's current state without a refresh. When it can't, it says so plainly instead of showing stale data as current.
- Honest under pressure
- In an incident the UI must not surprise you: no silent retries that start processes, no hidden destructive defaults, and Docker's errors shown as Docker wrote them.
- Weight is a feature
- Every addition earns its bundle size, memory and startup cost. Heavy tools load only when you open them.
- Auditable
- Every change leaves a record of who made it. Roles and audit are in the free edition.
- Built to grow past one host
- Environments, roles and the real-time channel were designed so agents and Kubernetes extend the product rather than fork it.
Where it is
| Core and local Docker | Shipped |
|---|---|
| Agent and multi-host | Shipped |
| Kubernetes and Helm | Shipped |
| Stacks and GitOps | Next |
| SSO, teams, API tokens, backup | Planned, Enterprise |
FleetOps is in early development. Follow along, or tell us what you need, on GitHub.