Administrator guide
This guide is for the people who run Vulnerability Operator: they install the server, connect the agents, decide who sees what, and keep the scans and alerts working. The User guide describes the day-to-day pages; the Reference lists every setting and endpoint.
The parts you run
| Part | Where it runs | Notes |
|---|---|---|
Server (vop-server) | a container or a Kubernetes deployment | Serves the API and the web interface; runs the scans and scheduled tasks. |
| Redis | a service you provide | Locks, caches, login throttling and the job queue. |
| S3-compatible storage | a service you provide | SBOM documents, generated reports and agent binaries. MinIO works. |
| Database | SQLite by default, or PostgreSQL, MySQL, MSSQL | SQLite lives in the /data volume of the server. |
| Web frontend | a container with nginx | Optional when the server is reached directly. |
| Linux, Windows and macOS agents | the hosts | A service that sends the inventory. |
| Kubernetes agents | every node of a cluster | A DaemonSet from its own Helm chart. |
First-day checklist
- Install the server. Use Docker Compose for a single machine, or the Helm chart for a cluster.
- Sign in as the initial administrator and change the password. If you did not
set
INITIAL_ADMIN_PASSWORD, the generated one is in the server log on first start. - Set the public address with
PUBLIC_URL, and enable TLS. See Security. - Check the scheduled tasks. The Grype vulnerability database is downloaded by
the
grype-db-updatetask; the first scan waits for it. See Scans and tasks. - Connect the first agents. Add hosts with Adding hosts, and clusters with Kubernetes agents.
- Create users and roles. Use the built-in viewer role for read-only access, and set scopes where teams should see only their clusters and nodes. See Users, roles and scopes.
- Configure single sign-on if your organisation uses an identity provider. See Single sign-on.
- Set up alerts and, if wanted, an AI backend for report summaries.
Platform status
Not every platform has been run end to end yet. The project's own status notes say:
- The server and the Linux agent were run end to end against Redis, an S3 endpoint and the real Grype database. PostgreSQL 16 was run for migrations, ingest, scans, filters, dashboard, package lists and PDF export.
- The Windows agent and its service have been built and released, but not yet run on a Windows host in this project's testing.
- The macOS agent has been built and released, but not yet run on a Mac.
- MySQL and MSSQL drivers are included but were not run.
- Delivery to Gotify and SMTP was not tested against real servers; ntfy was.
Test these in your own environment before you depend on them.