CnosDB's small web footprint still needs tenant and authentication review
CnosDB is a time-series database with a default HTTP API on port 8902 and a default tenant named cnosdb. The documentation says authentication is disabled by default through auth_enabled=false, the initial user is root, and the default root password is root in current documentation. Those defaults make a public service much more consequential than its small web footprint might suggest.
A count-only ZoomEye measurement on November 17, 2026 used sub_type=all, page 1, and a page size of 1. app="CnosDB" returned 0. title="CnosDB" returned 10. The application fingerprint produced no matches in this observation. The title set is small and suitable for a focused review, but a title match does not show whether the API is reachable or whether authentication is enabled.
What the counts can support
CnosDB's permissions guide is direct about the risk: if the service runs on a public endpoint, authentication should be enabled or data can be exposed to unauthenticated users. The guide recommends setting a root password, changing auth_enabled to true, and restarting the instance. It also describes tenants, users, roles, and admin permissions. The initial admin permission of the root user cannot be reclaimed, so root lifecycle needs special care.
The first review step is network placement. Determine which hosts listen on port 8902 and whether any public proxy forwards to that port. Check whether the service is exposed directly, through a management gateway, or only inside a private network. A title page may be a monitoring view while the API sits behind a different hostname. Record both.
The second step is authentication configuration. Confirm auth_enabled in the running configuration, not only in a deployment template. Check that the root password has been changed from the documented default. Inspect the must_change_password state for users that were created with temporary passwords. Review admin grants and remove accounts that do not need system authority.
The third step is tenant and role design. CnosDB uses tenants and roles to scope access. A single shared root account removes that boundary. Create service-specific users, grant least privilege, and document which tenant each application can reach. Review role inheritance and deletion behaviour. A stale membership can remain relevant after a role is removed, so test the actual access path.
The fourth step is transport and data exposure. The documentation notes TLS settings in the service configuration. Use encryption where the API crosses untrusted networks. Review what data the time-series service holds and whether it can contain infrastructure measurements, identity attributes, or application logs. A read-only endpoint can still be sensitive.
Detection should cover HTTP API access, authentication failures, user and tenant changes, and queries against system tables. Correlate network logs with CnosDB activity when the service is exposed through a proxy. A public endpoint that accepts requests without authentication is a high-priority configuration finding even if no exploit-related request appears.
The 0 application result and 10 title results should be retained together. The application fingerprint may not cover the current deployment style, and the title may be absent from a customised API response. The correct conclusion is that ten title candidates deserve ownership review, not that all are open databases.
Verification and next steps
CnosDB's small count makes a manual review feasible. Start with the exact defaults documented by the product, verify each owned service, and use network restrictions as a second layer. The necessary controls are a strong root password, authentication enabled, and a tenant model that matches the application.
Top comments (0)