Kubernetes Operating Model

Assistance-managed Kubernetes boundaries for customers and operators

Kubernetes is a platform for running containerized workloads with declarative deployment, service discovery, scaling, and recovery primitives. In Assistance engagements, Kubernetes is treated as an operated platform boundary: Assistance can design, build, improve, or run the cluster and platform services while your team owns the applications that run on it.

When to use Kubernetes#

Situation	Fit
Multiple services with repeatable deployments	Good fit when GitOps, templates, and shared platform controls reduce delivery risk
Existing clusters need reliability work	Good fit for Kubernetes Support or a managed platform plan
Moving from VMs or PaaS to more control	Good fit when workload readiness and team ownership are confirmed
Edge, lab, or constrained environments	Consider Managed Kubernetes
One small app with simple hosting needs	Kubernetes may add unnecessary operational complexity

What Assistance operates#

Platform area	Assistance responsibility
Cluster foundation	distribution selection, control plane, node pools, upgrades, capacity review, and runbooks
Networking	CNI, ingress, load balancers, DNS integration, TLS, and network policy baseline
Delivery	GitOps or CI/CD deployment path, Helm/Kustomize conventions, rollback workflow, and promotion controls
Security	RBAC, namespace model, pod security baseline, secrets integration, image policy guidance, and audit evidence where scoped
Observability	Prometheus/Grafana or provider metrics, logging, alerts, dashboards, and escalation routing
Reliability	etcd or cluster-state backup, restore notes, incident triage, maintenance windows, and post-incident notes

What the customer owns#

Area	Customer responsibility
Application workloads	code, manifests or chart values, health checks, resource behavior, and release correctness
Business data	classification, retention, migration acceptance, and application-level recovery validation
Access approvals	identity source, user approvals, repository permissions, and internal access reviews
Product impact	deployment timing, customer communications, and business decisions during incidents
Provider relationship	cloud account, billing, quotas, and contractual obligations unless Assistance-managed tenancy is contracted

Kubernetes is shared responsibility

Assistance can operate the platform, but application teams still own workload behavior. A healthy cluster cannot make an unsafe deployment, broken migration, or unbounded resource use correct without application ownership.

Engagement models#

Model	Best for
Readiness assessment	Decide whether Kubernetes is the right platform and what must change first
Platform build	Create a new cluster with GitOps, ingress, TLS, observability, security baseline, and handoff docs
Managed platform plan	Ongoing operations, upgrades, incident support, capacity review, and platform backlog management
Migration project	Move selected workloads into Kubernetes with staged rollout and rollback plans
Emergency response	Stabilize an active cluster or workload incident before ongoing support is established

Onboarding path#

Discovery — applications, dependencies, traffic, data, compliance, environments, deployment workflow, and team skills.
Target design — topology, distribution, networking, namespaces, GitOps, observability, secrets, backup, and support tier.
Build or takeover — provision a new cluster or baseline the existing one before Assistance accepts operational responsibility.
Workload onboarding — pilot service, deployment path, rollback plan, resource profile, and owner mapping.
Operate — upgrades, capacity reviews, incident support, maintenance windows, and platform improvements.

Normal change requests#

Use a planned request for new namespaces, ingress/DNS records, TLS changes, node pool changes, add-ons, policy changes, backup scope, or version upgrades. Emergency requests cover unavailable clusters, failed control plane operations, platform add-on outages, capacity exhaustion, and security events affecting covered services.

Getting started#

Start with a Kubernetes assessment. Assistance will map workload fit, operating boundaries, support expectations, and the safest path to build, improve, or migrate your platform.

Request Kubernetes assessment →