Operations: start here#
Running Whoobe: installing a workspace, configuring it, securing it, backing it up, upgrading it and watching it. Written for operators and DevOps.
The path for a new installation#
- Understand the shape. Architecture overview (what runs, ports, routing), and the deployment model: one workspace per organisation.
- Create a workspace. Creating workspaces is the tested install:
scripts/provision-workspace.sh, TLS, first administrator. - Configure. Configuration reference (every variable) and secrets and keys (what protects what, rotation).
- Protect. Encryption and encrypted backups; backup, restore and DR, scheduling backups; rehearse the restore once.
- Watch. Observability and logging.
- Size and scale. Capacity and scaling.
- Keep it current. Upgrade and migrations and the release process; read the release notes of the version (linked from the release section) before you upgrade.
All documents#
| Document | Purpose |
|---|---|
| Skills and effort | Who can run and maintain Whoobe: the profiles, the regular work and its estimated effort, what not to leave to one person |
| Creating workspaces | Create, secure, back up, restore and remove a workspace |
| Configuration reference | Every environment variable, by area |
| Secrets and keys | Inventory of secrets, lost or leaked, how to rotate |
| Encryption | At rest, in transit, backups, answers for a questionnaire |
| Backup, restore and DR | What to recover and a runbook |
| Scheduling backups | Nightly backup on the host, a copy kept on another machine, a check for a missed night |
| Audit log | Who changed what: what is recorded, how to read it, retention |
| Data subject requests and retention | Find, export and erase everything held about one person; delete by age |
| Single sign-on | Staff sign in with Keycloak, Entra, Okta or Google (OpenID Connect): set-up, roles, deprovisioning |
| Upgrade and migrations | Moving to the next release, going back |
| Capacity and scaling | What one workspace can do; what limits replicas |
| Observability and logging | What is recorded, what can be monitored, what is not available |
| Object storage | S3 stores that work |
| Private network | What the stack needs from the internet, the private profile and what is still open for an air-gapped site |
| Public sandbox | The restricted demo profile, rebuilt every three hours |
| Deployment guide (bare metal), miniguide | Native-service installs under PM2; not re-verified, the container workspace is the tested path |
Not covered yet#
Kubernetes manifests, a high-availability reference architecture, a managed-database compose variant (use DB_HOST and DB_SSL), rehearsed
runbooks for restore on a new host, upgrade and workspace re-creation.
Last updated