Nuclio is a serverless platform for deploying functions on Docker or Kubernetes. Its dashboard is not just a visual console. It exposes an HTTP API that reads and writes function configuration, including credentials carried by event triggers.
ZoomEye returned 446 matches for app="Nuclio" and 28 for title="Nuclio" on 2026-11-22 at 00:45 UTC. The query used sub_type=all, page 1, and pagesize=1. The application count is large enough to review as a population lead. It is not a count of vulnerable or internet-exposed dashboards.
Open the application query in ZoomEye. The title count is much smaller, which is common when a dashboard hides or changes its page title while the application fingerprint remains visible. The two fields describe different evidence.
The dashboard API is a credential boundary
Nuclio's production guidance warns that the dashboard API under /api/... reads and writes function configuration. That configuration can include top-level password and secret fields and nested values such as attributes.sasl.password or attributes.accesskey for event triggers.
The guidance says some managed platforms provide authentication for the dashboard. For deployments outside those platforms, operators should terminate authentication at the network layer with an authenticating ingress, an OAuth2 or OIDC proxy, or a service-mesh policy. It also says sensitive-field masking is controlled by platformConfig.sensitiveFields.maskSensitiveFields and is off by default in the official Helm chart.
An exposed dashboard API can therefore reveal more than function names. It can expose the authority to replace configuration, test functions, and read trigger credentials. The console port is often documented as 8070, but the actual service port and ingress depend on the deployment.
Confirm authentication at the actual edge
Do not stop at the presence of a login page. Test an unauthenticated request to the dashboard root, the API root, and a read-only function-list path from outside the approved network. Record the status code and response shape. Then test the ingresses, service mesh, and platform identity provider that should enforce the boundary.
Nuclio's documentation recommends network-layer authentication for non-managed deployments. That means the dashboard service should accept traffic only from the proxy or mesh identity, not from the broader internet. A network policy or private load balancer can complement application authentication, but it cannot protect a service that is intentionally published.
Review trigger secrets and function accounts separately. A dashboard bug is not required for exposure if an operator publishes the API without authentication. Rotate event-trigger credentials that were reachable through an unprotected dashboard and check function logs for unauthorized deployments or invocations.
Separate the function data plane
A function endpoint and the dashboard API are different services. A function may be public by design while the control plane is private. Measure them with separate queries, ports, and response checks.
The application count can help a security team find candidates for an inventory review, but it cannot tell whether a candidate is a dashboard, an API gateway, or a function runtime. Validate the response against Nuclio's documented paths and the operator's own DNS inventory.
For authorized hosts, compare the observed port with the deployment manifest. Confirm the dashboard service type, ingress annotations, service-mesh authorization policy, and any network-layer authentication. Check the sensitive-field masking setting and verify that secrets are redacted in the UI and API.
Build a repeatable baseline
Run the application and title queries independently at the same scope. Preserve the time, count, and query text. If the application count changes, investigate new deployments and exposure changes. If only the title count changes, consider custom dashboards and title noise before drawing a conclusion.
The useful outcome is not a list of 446 assumed victims. It is a reproducible candidate population for finding dashboards whose API can read or write function configuration. For each confirmed control plane, the next questions are who can reach it, which identity authenticates the request, and which trigger secrets it can reveal.