Architecture & cluster design
Cluster topology, node pools, networking, ingress and multi-environment layout.
Kubernetes · Consulting & Operations
Kubernetes Consulting and Managed Kubernetes Operations
MYC designs, builds, migrates and operates production Kubernetes on Huawei Cloud CCE or on clusters you run yourself. We don't sell a Kubernetes platform. We make the one in your account production-ready and keep it that way.
Who it's for
Kubernetes problems we are asked to fix
What we often find
- Clusters built by hand that nobody dares to upgrade.
- No resource requests or limits, so one workload starves the rest.
- Autoscaling missing or misconfigured, leaving capacity wasted or short at peak.
- Databases on Kubernetes without a plan for storage, backups or failover.
- Broad cluster-admin access and no network policies.
What we put in place
- Clusters defined as code with a tested upgrade path.
- Requests, limits and quotas sized from real usage.
- Pod and cluster autoscaling driven by the metrics that matter, such as queue lag.
- Operator-managed stateful workloads with persistent storage, backups and failover.
- RBAC, network policies and image scanning as defaults.
What we deliver
Kubernetes services from design to operations
Cluster build as code
CCE or self-managed clusters provisioned and configured from code, reproducible across environments.
Containerization & migration
Moving applications from VMs into Kubernetes, one service at a time.
Cloud migrationProduction readiness review
A structured review of reliability, scaling, security and operations, with prioritized fixes.
Stateful workloads
Databases and message brokers on Kubernetes with operators, StatefulSets and persistent volumes.
Autoscaling & cost
Horizontal pod autoscaling on custom metrics and right-sized node pools.
Security & multi-tenancy
RBAC, namespaces, network policies, secrets handling and hardened workload policies.
Delivery & GitOps
CI/CD into the cluster with image scanning, SBOMs and declarative deployments.
Platforms
Kubernetes tools we use in production
Kubernetes
Huawei Cloud CCE
Helm
Percona Operator
StatefulSets & CSI
HPA
Prometheus Adapter
NGINX Ingress
Prometheus
Grafana
Kafka / Redpanda
GitLab CI/CD
Trivy
Docker
How it works
How a Kubernetes engagement runs
-
01
Review
Assess current clusters or plans against production-readiness criteria.
-
02
Design
Target cluster architecture, tenancy model and delivery pipeline.
-
03
Build
Provision clusters and platform components from code.
-
04
Migrate
Move workloads in stages with health checks and rollback.
-
05
Operate
Upgrades, monitoring, incident response and capacity management.
-
06
Improve
Regular reviews of reliability, cost and developer experience.
Production readiness
What production-ready Kubernetes means to us
The checklist we review every cluster against.
Reliability
- Health probes and disruption budgets
- Spread across zones and nodes
- Tested backups for stateful data
Scaling
- Requests and limits from real usage
- Pod autoscaling on meaningful metrics
- Cluster autoscaling and quotas
Security
- Least-privilege RBAC
- Network policies between namespaces
- Scanned images and managed secrets
Operations
- Documented upgrade path
- Metrics, logs and alerts per workload
- Runbooks for common incidents
Proof
Kubernetes work we have delivered
Case study
Percona and Redpanda on Huawei Cloud CCE
Sharded and replica-set Percona clusters plus a three-broker Redpanda cluster, with workers autoscaling on Kafka lag.
Engineering
MongoDB on self-managed Kubernetes
StatefulSets, operator-managed replica sets and persistent volume tuning for production MongoDB.
Project
Kubernetes migration platform
Containerization, GitOps rollout and progressive cutover of legacy services.
FAQ
Kubernetes questions, answered
Do you provide a managed Kubernetes platform?
No. We design, build and operate Kubernetes in your own cloud account, usually Huawei Cloud CCE or a self-managed cluster. You keep ownership of the platform.
Should we use Huawei Cloud CCE or self-managed Kubernetes?
CCE removes control-plane operations and suits most teams on Huawei Cloud. Self-managed clusters make sense when you need a specific version, network setup or on-premise deployment. We recommend based on your constraints.
Can databases run on Kubernetes?
Yes, with operators, persistent storage, backups and tested failover. We run MongoDB (Percona) and Kafka-compatible brokers on Kubernetes in production.
Can you review a cluster we already run?
Yes. A production readiness review covers reliability, scaling, security and operations and ends with prioritized fixes.
How do you hand over to our team?
Everything is defined in code and documented, with runbooks for upgrades and common incidents.
Get Started
Request a Kubernetes architecture review
Tell us about your clusters or plans. We'll suggest what to review first.