🇵🇰 IAAS · COMPUTE · KUBERNETES CLUSTER SERVICE

Managed Kubernetes — for the teams shipping production inside Pakistan.

QCloud Kubernetes Cluster Service (KCS) is a CNCF-certified, fully managed Kubernetes platform. Spin up a production-grade cluster in under 10 minutes, scale to hundreds of nodes inside KHI-1 / ISB-1, and pay by the hour in PKR — without ever touching the control plane.

Launch your first cluster → Talk to a Kubernetes architect

CNCF-conformant · Tier-3 hosted · 99.999% multi-AZ SLA · Pay-as-you-go in PKR

cluster · prod-orders-khi-1 ● healthy
MANAGED BY QCLOUD Control Plane api-server · etcd · scheduler · controller-manager node-pool-app 3 nodes node-pool-batch 2 nodes node-pool-gpu 1 node · GPU
pod orders-api-7d4 scheduled → node-pool-app/2
Healthy ·6 nodes ·24 pods ·v1.30.2

Production clusters running on QCloud KCS today.

● live

Tier-1 Bank

core-payments

42 nodes · 380 pods · KHI-1 + ISB-1 · 99.999%

● live

National Telco

BSS-platform

28 nodes · 210 pods · 4 node pools · v1.30

● live

Listed FMCG

ecom-frontends

18 nodes · 96 pods · HPA + cluster-autoscaler

● live

Federal Agency

sovereign-stack

34 nodes · 240 pods · in-country only · PECA-aligned

🛡️ CNCF Conformant 🇵🇰 100% Data Residency ⚡ 99.999% SLA

One service. Two responsibilities. Zero servers to babysit.

QCloud KCS splits Kubernetes operations the way modern platform teams want: we run the control plane, you run the workloads. Everything in between is automated.

CONTROL PLANE

Managed by QCloud

  • ✓Multi-AZ Kubernetes API server (KHI-1 ↔ ISB-1)
  • ✓etcd cluster with automated backup + restore
  • ✓Scheduler, controller-manager, cloud-controller
  • ✓Automatic version upgrades on opt-in schedule
  • ✓CNCF Kubernetes Conformance — upstream APIs, no vendor lock-in
  • ✓Patched, hardened, monitored — 24/7 in-country SRE
WORKER NODES

Managed by you

  • ✓Node pools sized to your workload (CPU, memory, GPU)
  • ✓Your container images, your Helm charts, your manifests
  • ✓Your namespaces, your RBAC, your network policies
  • ✓Bring your own observability (Prometheus, Grafana, Datadog)
  • ✓GitOps / Terraform / kubectl — your tool, your rules
  • ✓Scale-in / scale-out via HPA + Cluster Autoscaler
kubectl · helm · argocd · flux · prometheus · grafana · istio · cilium · jaeger · kubeflow — all native, all supported

Four reasons enterprise teams pick QCloud KCS.

Built on the same Tier-3 platform that runs Pakistan's banks, telcos, and broadcasters — and tuned for the workloads container teams actually run.

Ease of Use — production K8s in 10 minutes

Automatic cluster creation through console, API, or Terraform. One-click rolling updates. One-stop deployment and Day-2 O&M. No "yaml-only" setup pain.

console · API · Terraform 1-click rolling update
01

High Performance — bare-metal speed

Kubernetes networking engineered from years of field experience. Auto-scaling that reacts in seconds. Bare-metal container performance — measurably faster than VM-based containers — at no extra cost.

bare-metal nodes <60s scale-out Cilium-class networking
02

High Availability — multi-AZ by default

Multi-AZ control plane, replicated etcd, automated hardware-failure recovery, graceful pod scale-out and scale-in. 99.999% multi-AZ SLA, no asterisks. Your workloads keep running when the hardware doesn't.

99.999% SLA multi-AZ etcd self-healing nodes
03

Open & CNCF-conformant — no lock-in, ever

Upstream Kubernetes (latest stable + N-1, N-2). CNCF-certified Conformance. Native Docker / containerd / CRI-O support. Built on the same APIs as EKS, GKE, AKS — port any workload in or out.

CNCF Certified upstream Kubernetes Docker · containerd
04

Pick the cluster shape that fits your team.

Three managed Kubernetes tiers — all CNCF-conformant, all multi-AZ-ready, all billed in PKR. Pick one to start, switch any time.

STARTER · DEV / STAGING

Managed Basic

Best for dev, staging, and small production workloads.

  • ✓Single-AZ Kubernetes control plane
  • ✓Up to 50 worker nodes
  • ✓Standard API request throughput
  • ✓Auto-upgrade on stable channel
  • ✓Audit logs · Prometheus-ready

From PKR ~12/hr per cluster · pay-as-you-go

Start Basic →
MOST POPULAR · PRODUCTION

Managed Pro

For production workloads with multi-AZ HA and tighter SLAs.

  • ✓Multi-AZ control plane (KHI-1 + ISB-1)
  • ✓Up to 1,000 worker nodes
  • ✓99.999% control-plane SLA
  • ✓Auto-upgrade with maintenance windows
  • ✓Advanced audit, encrypted etcd, KMS keys
  • ✓Premium SRE response: <15 min P1

From PKR ~42/hr per cluster · PAYG or Reserved

Start Pro →
ENTERPRISE · COMPLIANCE

Dedicated

Customer-managed control plane on dedicated QCloud hardware.

  • ✓Dedicated control-plane nodes — single-tenant
  • ✓Custom Kubernetes versions, custom add-ons
  • ✓Air-gapped option for sovereign workloads
  • ✓BYO etcd encryption keys (HSM-backed)
  • ✓Custom SLAs, custom node pool shapes
  • ✓Dedicated SRE pod, named contacts

Custom quote · annual commit

Talk to architect →
All three tiers run upstream Kubernetes. Switch between them without rewriting a single manifest.

Five capabilities. One control plane. No glue code.

KCS ships with the cloud-native primitives every production cluster needs — built in, not bolted on.

PRIMITIVE

Orchestration & Lifecycle

Schedule, deploy, restart, and roll out containerised workloads with full Kubernetes Deployment / StatefulSet / DaemonSet / Job / CronJob primitives. Rolling updates, blue/green, canary — out of the box.

apiVersion: apps/v1
kind: Deployment
metadata:
  name: orders-api
spec:
  replicas: 5
  template:
    spec:
      containers:
      - name: api
        image: orders:v4.13
Deployment StatefulSet DaemonSet Job CronJob

Auto-Scaling

HPA (Horizontal Pod Autoscaler), VPA, and Cluster Autoscaler — all preconfigured.

HPA VPA Cluster Autoscaler Karpenter-class

Cloud-Native Networking

Native VPC integration, in-cluster service mesh option (Istio / Cilium), LoadBalancer Services backed by QCloud ELB, Ingress controllers ready to install.

🛜 🔗 🚪 ⚖️
VPC-CNI Cilium Istio ELB Ingress

Persistent Storage

Block (EVS), object (OSS), and snapshot-backed PersistentVolumes. CSI driver included. Stateful workloads — Postgres, Kafka, ELK — run natively.

💽 🪣 📸
EVS-CSI OSS-CSI Snapshots BackupRestore

Security & Compliance

RBAC, Pod Security Standards, network policies, signed images, encrypted etcd, audit logs. Tier-3 hosted, PECA-aligned.

🛡️ ✓
RBAC PSS NetworkPolicy Signed Images Encrypted etcd

From git push to production pods — in one storyboard.

KCS plugs into the dev workflow your team already uses. Here is one developer journey, end to end.

01 · CODE
$ git commit -m "feat: invoice"

Engineers commit microservices to GitHub / GitLab / Bitbucket — KCS doesn't care which.

Git · any provider
02 · BUILD
building → invoice:v4.13

Build container images from Java, Python, Go, Node.js, or Vert.x source — push to QCloud Container Registry or any external registry.

Java · Python · Go · Node · Vert.x
03 · PUSH
→ 🗃️ ✓ signed

Signed, scanned, versioned. KCS pulls images only from registries you trust.

Image signing · CVE scan
04 · DEPLOY
kubectl rollout status

kubectl apply, Helm, ArgoCD, Flux — pick your delivery tool. Rolling updates, blue/green, canary supported natively.

kubectl · Helm · ArgoCD · Flux
05 · SCALE
pods 4 → 12 ▲

HPA scales pods on CPU / memory / custom metrics. Cluster Autoscaler scales nodes. Both work together, automatically.

HPA · VPA · CA
06 · OBSERVE

Prometheus metrics, Grafana dashboards, Loki logs, Jaeger traces — install in minutes, all CNCF-native.

Prometheus · Grafana · Loki · Jaeger
The whole loop runs inside Pakistan. Your code, your registry, your data — all in-country.

What teams actually run on QCloud KCS.

Six recurring workload patterns we see in production every week.

One cluster, hundreds of services, independent release cadences.

Run dozens to hundreds of independently-deployable microservices on one cluster. Each service scales horizontally, releases on its own cadence, and is discovered through native Kubernetes Service objects. Service mesh (Istio / Cilium) optional. Production patterns: API gateways, BFFs, event-driven workers.

IstioNGINX IngressgRPCgateway

GPU node pools, distributed training, model serving with Kubeflow.

Dedicate GPU-equipped node pools for training. Run TensorFlow / PyTorch jobs as Kubernetes Jobs or via Kubeflow pipelines. Serve models behind autoscaled inference Deployments. GitOps your model versions — KCS handles the rest.

KubeflowTensorFlowPyTorchGPU node poolsTriton

GitHub Actions runners, Jenkins agents, ArgoCD, Tekton, all in-cluster.

Run your build agents, test runners, and continuous-delivery controllers as Kubernetes workloads. Ephemeral runners scale to zero between jobs. Pull-based GitOps (ArgoCD / Flux) keeps cluster state synced to Git.

JenkinsGitHub ActionsArgoCDTektonFlux

Spark, Airflow, Flink — scheduled, parallel, scaled to zero overnight.

Schedule big-data jobs as Kubernetes CronJobs / Jobs. Run Spark-on-Kubernetes, Airflow workers, Flink JobManagers. Scale node pools up at job start, drain at job end — pay only for the runtime.

SparkAirflowFlinkArgo Workflows

Per-tenant namespaces, RBAC, quotas, network policies.

Isolate customers in dedicated namespaces with ResourceQuotas and NetworkPolicies. Common control plane, per-tenant data plane. Ideal for in-country SaaS vendors serving banks, telcos, and government.

NamespacesResourceQuotaNetworkPolicyvCluster

Lift VM workloads onto containers, in stages.

Containerise a single service from your monolith, run it on KCS alongside the rest. Iterate. Migration team helps you decompose, containerise, and cut over without big-bang risk.

Strangler-figSidecarHybridMigration support

KCS vs running Kubernetes yourself.

Kubernetes is open source — yes. But "free" isn't the same as "no cost". Here's what changes when QCloud runs the control plane for you.

SELF-MANAGED vs QCLOUD KCS
Stand up etcd, schedulers, API servers — yourself
Multi-AZ control plane — managed by QCloud SRE
Hire dedicated Kubernetes SREs (2–4 FTEs)
Use your existing platform team — KCS handles Day-2
Upgrade-night marathons every 3 months
Opt-in maintenance windows — KCS upgrades for you
Patch CVEs across control + worker plane
Control-plane CVEs patched by QCloud, 24/7
Build your own etcd backup / restore tooling
etcd auto-backup with restore-to-point in <1 click
Pay for control-plane VMs even when idle
Pay per cluster-hour in PKR — no idle tax
DIY observability for the control plane
Native control-plane logs into Prometheus + audit
Compliance evidence: yours to gather
Tier-3, CNCF, PECA-aligned — evidence on tap
Most teams reclaim 2–4 SRE FTEs and 30–40% of their compute spend within the first quarter.

The Kubernetes ecosystem — already plugged in.

KCS speaks upstream Kubernetes. That means every CNCF graduate and incubating tool you care about works out of the box. Here are the ones our customers ship with most often.

Kubeflow
OpenTelemetry
cert-manager
external-dns
Velero
Linkerd
Knative
Tekton
Falco
OPA / Gatekeeper
Helm
Prometheus
Grafana
ArgoCD
Flux
Istio
Cilium
Jaeger
K8s
upstream · v1.30
Plus — Helm chart registry, Terraform provider, GitOps support out of the box.

From apply to Running — five steps, ten minutes.

Click any step to see what runs on the right.

~/qcloud — kcs · create cluster 
qcloud@laptop ~ $ qcloud kcs cluster create \
  --name prod-orders \
  --region khi-1,isb-1 \
  --tier managed-pro \
  --version 1.30.2 \
  --vpc vpc-prod-01

✓ control plane provisioned (multi-AZ)
✓ etcd cluster healthy
✓ api server endpoint: https://k8s-prod-orders.qcloud.pk
✓ kubeconfig downloaded

cluster prod-orders → Running   (8m 42s)▋

Numbers we publish — and stand behind.

Real production data from QCloud KCS clusters in KHI-1 and ISB-1 over the last 12 months.

CONTROL PLANE AVAILABILITY

0%

Across multi-AZ Managed Pro clusters · last 12 months

MEDIAN POD SCHEDULING LATENCY

< 0s

From kubectl apply → pod Running · p50

CLUSTER PROVISIONING — END TO END

< 0 min

From "create" click → first workload scheduled

Sources: QCloud KCS production telemetry · KHI-1 & ISB-1 · 12-month rolling window.

Everything you wanted to ask about KCS, in one place.

If your question isn't here, ask a Kubernetes architect — we answer within the same business day.

KCS is QCloud's managed, CNCF-conformant Kubernetes platform hosted inside Pakistan. We run the control plane (API server, etcd, scheduler, controller-manager) across multi-AZ infrastructure in KHI-1 and ISB-1 — you bring your workloads. Everything in between (upgrades, etcd backup, certificate rotation, audit) is automated.
Yes. KCS runs upstream Kubernetes and is conformant with the CNCF Kubernetes Conformance Program. Any workload that runs on EKS, GKE, or AKS will run on KCS without modification — same APIs, same primitives, same kubectl.
A Managed Basic cluster is typically Running in under 8 minutes. A Managed Pro (multi-AZ) cluster takes 10–12 minutes end-to-end. Most customers go from signup → KYC → first workload scheduled in under one business day.
We support the latest stable Kubernetes minor version plus the two preceding minors (currently 1.30 / 1.29 / 1.28). Auto-upgrade is opt-in, with maintenance windows you choose. Manual upgrade through the console or API is one click.
Pay-as-you-go by the cluster-hour, in PKR. Managed Basic from ~PKR 12/hr, Managed Pro from ~PKR 42/hr. Worker nodes are billed separately at standard ECS-GP rates. 1- and 3-year Reserved discounts available for steady workloads. No FX exposure.
Yes. You can attach GPU-equipped node pools to any cluster. KCS supports the standard NVIDIA device plugin for Kubernetes and runs Kubeflow, TensorFlow, PyTorch, and Triton without modification. Distributed training across multiple GPU nodes is supported.
Managed Pro and Dedicated clusters run the Kubernetes control plane across multiple availability zones (KHI-1 and ISB-1). etcd is replicated across AZs. Worker node pools can also span AZs. The published control-plane SLA is 99.999%. Self-healing replaces unhealthy nodes automatically.
Yes — 100%. Every byte of cluster state, every container image (when stored in QCloud Container Registry), every PersistentVolume, every etcd backup, every audit log stays in PTCL's Tier-3 / Rated-3 certified data centers in Karachi (KHI-1) and Islamabad (ISB-1). No data ever leaves the country.
🇵🇰 PAKISTAN'S MANAGED KUBERNETES

Spin up your first KCS cluster — on the house.

$100 in QCloud credit. A 1-month free trial of Managed Pro. A real Kubernetes architect on the call from day one. Migration support for teams moving off self-managed K8s, EKS, GKE, or AKS — all included.

Take a demo → Or talk to cloud sales

No credit card required · Production workloads welcome · PKR billing

● Live · 9 clusters provisioned across QCloud KCS this week

📞 Call cloud sales · 021 111 486 583 ✉️ Email · info@qcloud.pk
console › kcs › create cluster
prod-orders-khi
KHI-1 ISB-1
Basic Pro Dedicated
1.30.2
gp.m4.large × 3 nodes
vpc-prod-01 / subnet-private-a,b
▶ Create cluster

Provisioning typically completes in 8–12 minutes.

Scroll to Top