Skip to main content

Security

This section answers the question an operator actually has: what does OLIVE support, and what do I have to provide for it?

OLIVE is delivered as an artefact — microservices, Helm charts, container images, configuration defaults and the pipeline that builds them. It operates nothing. Several of its security controls therefore depend on capabilities the cluster provides: a CNI that enforces NetworkPolicy, an Istio ambient installation, a Kyverno admission controller, a log platform, backup infrastructure.

The short version: the platform makes the control possible; the cluster makes it work; the platform administrator switches it on. If any of the three is missing, the control is not in place — and in most cases nothing will tell you.

Pages in this section​

PageAnswers
Security CapabilitiesWhat OLIVE supports out of the box, what is switchable, and what the cluster must provide for each
Service Mesh and mTLSIstio ambient mode: why it was chosen, what it needs, how it interacts with NetworkPolicy
Admission Control and Image IntegrityKyverno, cosign signatures, digest pinning, SBOM, registry allow-list
Workload HardeningWhat the charts set on every pod and why, plus Pod Security Admission
Shared ResponsibilityWho owns which control: OLIVE platform team, cluster operator, platform administrator, customizer
What OLIVE requires of Vault, Keycloak, Istio and Kyverno

Each of these has documented requirements — the exact Vault bootstrap policy, the Kyverno label contract, the mesh exemptions — and much of meeting them is your work rather than the platform administrator's. That contract is part of the platform; install the components however you like.

Setup charts for all of them are also available, as-is and outside the platform's support and maintenance. See Infrastructure Charts.

Related pages elsewhere:

What OLIVE does not do​

Stated plainly, because the boundary matters more than any individual feature:

  • OLIVE does not detect or respond to incidents. It emits the signals — a defined catalogue of security events with stable codes, in machine-readable form. Alert rules, triage, escalation, containment and statutory reporting are operational capabilities you build on top. Anomaly detection, per-tenant baselines and impossible-travel analysis are deliberately not built into the product rather than built badly.
  • OLIVE does not harden your cluster. Node configuration, the CNI, cloud IAM, perimeter firewalls, the monitoring stack and encryption at rest for volumes, backups, object storage and etcd are yours.
  • OLIVE does not back itself up. It documents what must be preserved — including, for every Kubernetes Secret, where its authoritative copy lives and what happens if it disappears — but backup coverage, scheduling, isolation, restore testing and RPO/RTO are yours. See What to Back Up.
  • OLIVE does not scan uploaded files for malware. File type, extension and MIME validation and size limits are in place; no virus scanner is applied.