Built for the teams that run the hosts.

Five situations FleetOps was designed around. Each is the same product; what changes is the part you lean on.

Shared Docker hosts

Small teams where several people deploy to the same daemon.

The problem

Everyone has SSH and root-equivalent socket access, nobody knows who restarted what, and the only dashboard is docker ps.

With FleetOps

Give each person a login and a role instead of a shell. Viewers read, operators change things, and every change is on the record.

Roles per environment: Viewer, Operator and Admin, granted per host.
Audit log: Who did what, including every terminal session and every refused request.
No SSH needed: Logs, terminal, stats and files from the browser.

Teams that run Docker hosts and Kubernetes clusters side by side.

The problem

Docker in one tool, clusters in another, Helm in a third, and a different mental model in each.

With FleetOps

One server, one UI and one set of roles across every environment, each carrying its own colour band and tag so you always know where you are.

Kubernetes: Workloads, services, config, storage, RBAC and custom resources, on live watches.
Safe YAML edits: Every write is a server-side dry run, shown as a diff first.
Helm: Install, upgrade, roll back, each reviewed object by object.

The engineer paged at 2 a.m., beside a terminal and an incident channel.

The problem

Stale dashboards, a log tail that freezes at 10,000 lines, and a shell that quietly reconnects to a new container.

With FleetOps

The failing container rises to the top on its own. Walk to it with j and k, read its logs, recreate it, and the record writes itself.

Needs attention: Exited, restarting and unhealthy containers, racked first.
Log viewer: 200,000 lines, regex search, resume without repeats.
Honest terminal: A dropped session ends; it never respawns a shell behind your back.

Hosts in stores, labs, factories or a second cloud, often behind NAT.

The problem

Opening a Docker port to reach a remote host is a risk, and a VPN per site is a project.

With FleetOps

Run the agent on the host. It dials out to the server over one connection, so the host opens no port and never exposes its socket.

Dial-out agent: Outbound only, through NAT and firewalls.
One-time join: A 15-minute, single-use token, then rotate or revoke at will.
Same tools remotely: Logs, terminal, stats and image builds work as they do locally.

Teams whose servers never see the internet, and auditors who ask for evidence.

The problem

Tools that need a license server, phone home for telemetry, or keep audit trails you cannot take out.

With FleetOps

FleetOps is one binary on your hardware. Licenses are signed files checked offline, and the audit log exports to your SIEM.

Offline license: No license server, no call home, a 14-day grace after expiry.
Audit export: CSV or NDJSON over any date range, on paid plans.
Small surface: Distroless image, runs as non-root, strict CSP.

Community is free on any number of nodes, with roles and audit included.