What we do
- Configuration: a hardened baseline for each kind of machine, written as Ansible playbooks and kept in your Git repository: users and SSH keys, updates, time and DNS, firewall, VPN, Docker and the software each role needs. A new machine is ready in minutes, set up exactly like the others
- Management: patching, software and setting changes rolled out to every machine at once, or one group at a time; dry runs that show what would change before anything does; reports of machines that have drifted from the baseline
- Monitoring, set up by Ansible: exporters installed on every machine, added to Prometheus, alert rules for disks, memory, temperature, hardware health and failed backups, dashboards in Grafana. A new machine appears on the dashboards as soon as it is built
- Support: playbook reviews and fixes, failed runs looked into, upgrades of Ansible and its collections, and a web console (Semaphore UI) so your team can run approved jobs without the command line
- Mixed estates: Linux over SSH; Windows over OpenSSH or WinRM; VMware vSphere and ESXi through their API; Dell servers through iDRAC (Redfish), and older ones through IPMI
From a change to every machine
Text version of this diagram
- Change to a playbook → Review in Git
- Review in Git → Dry run: what would change
- Dry run: what would change → Approved run from the web console
- Approved run from the web console → Linux, Windows, VMware, Dell iDRAC
- Linux, Windows, VMware, Dell iDRAC → Prometheus and Grafana
- Prometheus and Grafana → Engineer looks into it (alert)
Versions we work with
- Ansible: ansible-core 2.21 and the current Ansible community package, with Semaphore UI as the web console
- Linux: Ubuntu 26.04 LTS and 24.04 LTS, Debian 13, Red Hat Enterprise Linux 10 and compatible distributions, Raspberry Pi OS
- Windows: Windows Server 2025, Windows 11
- Virtualization: VMware vSphere 9 and 8, Proxmox VE 9
- Hardware: Dell iDRAC 10 and 9
- Monitoring: Prometheus 3, Grafana 13
- Older versions: managed as they are, with a plan to bring them up to date
Typical projects
- A first baseline: an inventory of every machine, one playbook per role, a dry run against what's running today, then the differences settled one by one
- New machines from bare hardware to monitored and backed up in one run
- A web console for the team, with access by role and every run logged
Ansible is a trademark of its owner. ServerWorks is an independent service provider and is not affiliated with or endorsed by its owner.
Questions
Is Ansible free?
Yes: Ansible is open source. Red Hat sells Ansible Automation Platform with support; most estates don't need it, and we work with either.
Does Ansible need an agent on every machine?
No: it connects over SSH (or WinRM on Windows) when a job runs, and leaves nothing running.
Will a mistake reach every machine at once?
Not unreviewed: changes are reviewed in Git, shown in a dry run, and rolled out to one group first.
Where are passwords kept?
Encrypted with Ansible Vault, never in plain text in the repository.