Every cluster category, and where we actually stand
Vendors answer "what do you do?" with a logo wall. We answer it with an index — 22 domains, 233 cluster categories, each one marked with the depth of our practice. The categories marked core are what we do every week. The ones marked not ours are where we will send you to someone better rather than learn on your budget.
Five states, and no unmarked entries
Every category carries one of five marks. There is no "we do a bit of everything" entry, because that phrase carries no information a buyer can act on.
- Core practice 137 We design, build, harden and operate this ourselves. It is what we do every week.
- Regular delivery 57 Delivered repeatedly as part of standard engagements. Reference designs exist.
- Advisory 23 We assess and design, hand over the runbook, and work alongside your team.
- Partner-led 11 We integrate and coordinate; the vendor or a specialist partner owns the platform.
- Not ours 5 Outside our lane. We will tell you who to call instead of learning on your budget.
Container orchestration & platform
The layer that runs everything else as a workload. Distribution choice matters far less than whether you can upgrade and recover it.
- Core practiceour practice
Kubernetes — self-managed HA control plane
kubeadm · etcd quorum · containerd · HAProxy/kube-vip
- Core practice
Kubernetes — lightweight & edge distributions
K3s · k0s · MicroK8s · Talos Linux
- Core practice
Kubernetes — enterprise platform distributions
RKE2 · OpenShift / OKD · Rancher · D2iQ
- Core practice
Kubernetes — declarative cluster lifecycle
Cluster API · Crossplane · Talos machine config
- Core practice
Kubernetes — bare-metal load balancing
MetalLB (L2/BGP) · kube-vip · PureLB
- Core practice
Kubernetes — multi-tenancy & policy
Kyverno · OPA Gatekeeper · vCluster · Capsule
- Core practice
Kubernetes — ingress & Gateway API
Gateway API · ingress-nginx · Envoy Gateway · Traefik
- Core practice
Kubernetes — storage & CSI
Rook-Ceph · Longhorn · local-path · CSI drivers
- Regular delivery
Kubernetes — autoscaling & scheduling
HPA / VPA · KEDA · Karpenter · Cluster Autoscaler
- Regular delivery
Kubernetes — service mesh
Istio · Linkerd · Cilium Service Mesh · Consul Connect
- Advisory
Nomad — workload orchestration
HashiCorp Nomad · Consul · Vault
- Advisory
Docker Swarm — legacy estates
Swarm mode · overlay networks
Virtualisation, hyperconvergence & VDI
Where most refresh budgets currently sit. The migration question is no longer whether, but in how many waves.
- Core practice
VMware → open-source migration
ESXi → Proxmox VE + Ceph · KubeVirt · virt-v2v
- Core practice
Proxmox VE cluster
PVE 8 · Corosync · Ceph · live migration
- Core practice
Harvester hyperconverged
Harvester · Longhorn · KubeVirt
- Core practice
KubeVirt — VMs on Kubernetes
KubeVirt · CDI · KubeVirt operator
- Core practiceour practice
VDI — open source / self-hosted
Proxmox VDI · Kasm Workspaces · Guacamole · X2Go
- Regular deliveryour practice
VMware vSphere HA / DRS cluster
vSphere 8 · vMotion · DRS · vSAN
- Regular delivery
OpenStack private cloud
Kolla-Ansible · Nova · Neutron · Cinder · Ceph
- Regular delivery
Application publishing / virtual apps
Kasm · Citrix · RDS Session Host farm
- Advisory
XCP-ng / XenServer resource pool
XCP-ng 8 · Xen Orchestra
- Advisory
oVirt / Red Hat Virtualization
oVirt 4 · RHV
- Advisory
VDI — VMware Horizon
Horizon 8 · UAG · instant clones · App Volumes
- Advisory
VDI — Citrix Virtual Apps & Desktops
CVAD · NetScaler · MCS/PVS
- Partner-led
Nutanix AHV cluster
AHV · Prism Central · AOS
- Not ours
Mainframe & POWER virtualisation
z/VM · PowerVM · LPAR · HMC
High availability & failover
The discipline underneath everything else. An HA cluster that has never failed over on purpose is an untested assertion.
- Core practiceour practice
Pacemaker + Corosync cluster
pcs / crmsh · SBD fencing · STONITH · qdevice
- Core practice
Keepalived / VRRP failover
keepalived · VRRP instance · track_script
- Core practice
Block replication HA
DRBD 9 · LINSTOR · dual-primary
- Core practice
Floating-VIP application HA
systemd units · notify · ipaddr2 resource
- Core practice
Load-balancer HA pair
LVS + keepalived · HAProxy pair · VIP handoff
- Core practice
Sharded / active-active application tier
consistent hashing · quorum · split-brain guards
- Advisory
Windows Server Failover Cluster
WSFC · SQL AlwaysOn · Storage Spaces Direct
- Partner-led
Veritas InfoScale / VCS
InfoScale · CFS · VCS agents
- Partner-led
SIOS LifeKeeper
LifeKeeper for Linux
HPC & batch scheduling
Where queue time, not hardware, is usually the bottleneck. Scheduler policy is a measurable deliverable.
- Core practiceour practice
Slurm cluster
Slurm 24.x · Munge · cgroup v2 · accounting (slurmdbd)
- Core practice
Slurm on Kubernetes
Slinky · Slurm Operator · Kueue integration
- Core practice
Batch scheduling on Kubernetes
Volcano · Kueue · JobSet · gang scheduling
- Core practice
Ray cluster
Ray · Ray Serve · KubeRay · autoscaler
- Regular delivery
Apache Spark cluster
Spark 4 · on K8s / YARN · dynamic allocation
- Advisory
PBS Pro / OpenPBS / Torque
OpenPBS · PBS Professional
- Advisory
SGE / Grid Engine lineage
SGE · Son of Grid Engine · Altair Grid Engine
- Advisory
Render & simulation farm
Thinkbox Deadline · OpenCue · burst to cloud
- Partner-led
IBM Spectrum LSF
LSF · Spectrum Cluster Foundation
GPU & AI infrastructure
The fastest-moving category on this page, and the one where utilisation reporting separates engineers from resellers.
- Core practiceour practice
GPU training cluster
NVIDIA GPU Operator · CUDA · MIG · DCGM
- Core practice
GPU scheduling & sharing
DRA · MPS · time-slicing · Kueue · gang scheduling
- Core practice
Accelerator fabric
InfiniBand · RoCE v2 · NCCL · GPUDirect RDMA
- Core practice
Parallel storage for AI
BeeGFS · Lustre · WekaFS · IBM Storage Scale
- Core practice
LLM inference cluster
vLLM · Triton · KServe · llm-d · KV-cache routing
- Core practice
GPU cost & utilisation reporting
DCGM exporter · OpenCost · per-team chargeback
- Regular delivery
Training orchestration
Kubeflow · MLflow · Ray Train · PyTorchJob
- Regular delivery
Vector database cluster
Milvus · Qdrant · pgvector · Weaviate
- Regular delivery
Model / inference gateway
LiteLLM · Envoy AI Gateway · token accounting
Relational & distributed SQL
Our deepest practice. Every design decision here is about one thing: can you fail over and restore, and do you know in advance how long it takes.
- Core practiceour practice
PostgreSQL high availability
Patroni · etcd · pgBackRest · PgBouncer
- Core practice
PostgreSQL synchronous replication cluster
repmgr · sync standby · quorum commit
- Core practice
MySQL / MariaDB HA
InnoDB Cluster · Group Replication · MySQL Router
- Core practice
Galera cluster
MariaDB Galera 4 · Percona XtraDB Cluster · garbd
- Core practice
Distributed SQL — TiDB
TiDB · TiKV · PD · TiFlash
- Core practice
Distributed SQL — CockroachDB
CockroachDB · multi-region · geo-partitioning
- Core practice
Database proxy & pooling layer
ProxySQL · PgBouncer · MaxScale · ShardingSphere
- Core practice
Database migration & replication pipelines
pglogical · Debezium · Ora2Pg · mydumper
- Core practice
Zero-downtime major version upgrade
logical replication cutover · blue/green
- Regular delivery
MySQL orchestration & failover
orchestrator · MHA · ProxySQL
- Regular delivery
Distributed SQL — YugabyteDB
YSQL · DocDB · RAFT groups
- Regular delivery
Distributed PostgreSQL — Citus
Citus · shard rebalancing
- Advisory
Distributed SQL — OceanBase / PolarDB-X
vendor distributions
- Advisory
MPP analytics database
Greenplum · Cloudberry · Vertica
- Advisory
SQL Server AlwaysOn
Availability Groups · FCI · DAG
- Partner-led
Oracle RAC / Data Guard
RAC · Active Data Guard · Grid Infrastructure
- Not ours
Mainframe databases
Db2 for z/OS · IMS · VSAM
NoSQL, cache, search & coordination
The clusters that quietly become single points of failure because nobody calls them "the database".
- Core practice
MongoDB replica set
MongoDB 8 · write concern · oplog · PSA
- Core practice
MongoDB sharded cluster
mongos · config server replica set · balancer
- Core practice
Redis Cluster / Sentinel
Redis 7 · Valkey · cluster mode · ACL
- Core practice
etcd cluster
etcd 3.5 · raft · defrag · snapshot restore
- Core practice
Search cluster
Elasticsearch 8/9 · OpenSearch · shard & ILM planning
- Core practice
Time-series store
TimescaleDB · InfluxDB · VictoriaMetrics
- Regular delivery
Wide-column ring
Apache Cassandra · ScyllaDB · nodetool repair
- Regular delivery
Memcached tier
memcached · consistent hashing client-side
- Regular delivery
ZooKeeper ensemble
Apache ZooKeeper · dynamic reconfig
- Advisory
Solr cluster
Apache Solr · SolrCloud · ZooKeeper
- Advisory
Graph database cluster
Neo4j causal cluster · routing drivers
Messaging, streaming & events
Ordering, durability and consumer lag are the only three questions that matter. Everything else is configuration.
- Core practiceour practice
Kafka cluster (KRaft)
Kafka 4 · KRaft · MirrorMaker 2 · tiered storage
- Core practice
NATS JetStream
NATS · JetStream · leaf nodes · super-clusters
- Core practice
RabbitMQ cluster
quorum queues · streams · Shovel/Federation
- Core practice
Change data capture
Debezium · Kafka Connect · outbox pattern
- Regular delivery
Kafka — legacy ZooKeeper mode
Kafka 3.x · ZooKeeper ensemble
- Regular delivery
Apache Pulsar cluster
Pulsar · BookKeeper · tiered storage
- Regular delivery
ActiveMQ Artemis
Artemis · HA replication pair
- Regular delivery
MQTT broker cluster
EMQX · HiveMQ · VerneMQ
- Regular delivery
Schema registry & governance
Apicurio · Confluent Schema Registry
- Advisory
Apache RocketMQ
RocketMQ · DLedger
Analytics & lakehouse
The line between "data platform" and "cluster engineering" is storage layout and query planning. We work on the second.
- Core practiceour practice
ClickHouse cluster
ClickHouse · Keeper · shards & replicas · distributed tables
- Core practice
Trino / Presto cluster
Trino · coordinator/worker · fault-tolerant execution
- Core practice
Apache Flink cluster
Flink · Kubernetes Operator · HA · savepoints
- Core practice
Lakehouse table formats
Apache Iceberg · Delta Lake · Hudi
- Core practice
Object-store lakehouse
S3-compatible · MinIO · lakeFS · compaction jobs
- Core practice
Stream–batch unification
Flink + Iceberg + Trino · CDC to lake
- Regular delivery
Real-time OLAP
StarRocks · Apache Doris · Pinot · Druid
- Regular delivery
Hadoop HDFS / YARN HA
NameNode HA · JournalNode · ResourceManager HA
- Regular delivery
Batch ETL orchestration
Airflow · Dagster · dbt
- Regular delivery
Analytics query gateway
Trino as federation layer · row/column policy
Distributed & parallel storage
Performance problems are usually placement problems. CRUSH maps, failure domains and rebalancing windows explain more than any benchmark.
- Core practiceour practice
Ceph cluster
Ceph Reef/Squid · MON/MGR/OSD · CRUSH · BlueStore
- Core practice
CephFS & RBD
MDS active/standby · RBD mirroring · snapshots
- Core practice
Rook-Ceph on Kubernetes
Rook operator · OSD on PVC/LV · CSI
- Core practice
S3-compatible object storage
MinIO · erasure coding · site replication
- Core practice
Longhorn
Longhorn · replica rebuild · snapshots
- Core practice
Parallel filesystem (HPC)
BeeGFS · Lustre · WekaFS · IBM Storage Scale
- Core practice
Scale-out NAS
NFS-Ganesha · clustered Samba · pNFS
- Core practice
Storage QoS & tiering
Ceph QoS · NVMe cache tier · placement rules
- Core practice
Stretch cluster & replication
RBD mirroring · SnapMirror · async/sync policy
- Core practice
Capacity & growth modelling
CRUSH redistribution forecast · rebalance windowing
- Core practice
Storage migration between platforms
rsync-over-cluster · rclone · block copy with cutover
- Regular delivery
SeaweedFS / JuiceFS / MooseFS
S3 gateway · metadata engines
- Regular delivery
OpenStack Swift
Swift · proxy/account/container rings
- Regular delivery
HDFS as storage tier
DataNode placement · rack awareness
- Advisory
GlusterFS — legacy estates
Gluster 11 · arbiter volumes
File, block & object services
The services users actually touch. Nobody notices them until permissions, locking or path failover breaks.
- Core practiceour practice
SMB / Samba scale-out cluster
clustered Samba · CTDB · ctdb public addresses
- Core practice
NFS server HA
NFS-Ganesha · DRBD-backed export · VIP failover
- Core practice
S3 front door / gateway
Ceph RGW · MinIO gateway · bucket policies
- Core practice
Block storage over network
iSCSI (LIO/targetcli) · NVMe-oF TCP/RDMA
- Core practice
Quota, audit & access governance
POSIX ACLs · Windows ACL mapping · auditd
- Core practice
Home directory at scale
NFS/SMB + AD integration · profile roaming
- Regular delivery
Immutable / WORM archive tier
S3 Object Lock · retention policies
- Advisory
Windows file services & DFS
DFS-R · DFS-N · Scale-Out File Server
- Partner-led
SAN array integration
Dell PowerStore/PowerMax · Pure · NetApp · HPE
Backup, DR & data protection
The only category on this page where the deliverable is a number someone measured. "We have backups" is not a recovery plan.
- Core practiceour practice
Self-hosted backup cluster
Bacula · Bareos · Amanda · Restic/Borg + repo
- Core practice
Proxmox Backup Server
PBS · deduplication · sync & verify jobs
- Core practice
Kubernetes backup & restore
Velero · Kasten K10 · CSI snapshots
- Core practice
Database backup at scale
pgBackRest · XtraBackup · mongodump/mongorestore
- Core practice
Off-site & air-gapped vault
3-2-1-1-0 · immutable repo · Object Lock
- Core practice
Dual-site active/active & arbitration
stretched cluster · witness arbiter · quorum device
- Core practice
DR orchestration & runbooks
Velero · scripted failover · DNS cutover
- Core practice
Restore drills & verification
timed restore · checksum verification · report
- Regular delivery
Ransomware-resilient backup design
hardened repo · offline copy · immutability
- Partner-led
Backup appliance integration
Veeam · Rubrik · Cohesity · Data Domain
Network devices & fabric
Devices are not clusters until someone makes them redundant on purpose. A pair sharing one PDU is one device with extra plastic.
- Core practiceour practice
Firewall HA cluster
Fortinet FGCP · PAN-OS HA3 · Check Point ClusterXL · CARP/pfsync
- Core practice
Switch stack / MLAG
MLAG · MC-LAG · vPC · VSS · stack cabling
- Core practice
EVPN-VXLAN spine-leaf
Arista EOS · Cumulus · SONiC · FRR
- Core practice
Routing overlay & fast convergence
FRR · BGP · OSPF · IS-IS · BFD · VRF
- Core practice
Network automation & source of truth
NetBox · Jinja2 templating · Oxidized · NAPALM
- Core practice
Network validation & simulation
containerlab · Batfish · pre-change analysis
- Core practice
Network telemetry
sFlow · gNMI · SNMP · LibreNMS · LibreQoS
- Core practice
Out-of-band management plane
serial console servers · segregated OOB · jump host
- Regular delivery
Firewall platform migration
policy conversion · rule clean-up · shadowed-rule analysis
- Regular delivery
Routing security
RPKI · prefix filters · uRPF · max-prefix guards
- Regular delivery
Network access control
802.1X · FreeRADIUS · NAC policy
- Partner-led
Structured cabling & optical layer
MPO/LC · DWDM · patch documentation
Load balancing, ingress & mesh
Traffic entry is where availability is either delivered or quietly given away. Health checks that always pass are worse than none.
- Core practiceour practice
L4 load balancing
LVS/IPVS · DSR mode · Maglev · ECMP
- Core practice
L7 load balancing
HAProxy · nginx · Envoy · active/passive pair
- Core practice
API gateway
Kong · APISIX · Envoy Gateway · rate limiting
- Core practice
TLS termination & certificate lifecycle
ACME/Let's Encrypt · cert-manager · internal CA · mTLS
- Core practice
Container ingress
Envoy Gateway · Traefik · MetalLB BGP
- Regular delivery
Global server load balancing
GSLB · DNS steering · health-based failover
- Regular delivery
Service mesh
Istio · Linkerd · Cilium Service Mesh
- Regular delivery
WAF cluster
ModSecurity/CRS · reverse-proxy WAF · tuning
- Advisory
DDoS mitigation & abuse controls
scrubbing · rate limiting · SYN cookies
- Partner-led
Hardware ADC
F5 BIG-IP · Citrix ADC · A10 Thunder
Foundation network services
DNS, DHCP and time. Nobody plans capacity for them, and when they fail the whole company files a ticket in the same minute.
- Core practiceour practice
Anycast DNS cluster
BIND 9 · Knot DNS · PowerDNS · Unbound
- Core practice
DNS HA & hidden primary
secondary zones · catalog zones · TSIG · DNSSEC
- Core practice
DHCP HA
Kea HA hook · ISC DHCP failover · subnet delegation
- Core practice
Time & frequency services
chrony · ntpd · GPS/GNSS refclock · PTP grandmaster
- Core practice
Internal service discovery
CoreDNS · Consul DNS · external-dns
- Core practice
Certificate authority cluster
step-ca · EJBCA · HSM-backed root · ACME
- Regular delivery
AAA services
FreeRADIUS (HA) · tac_plus · AD integration
- Regular delivery
Mail transfer & relay cluster
Postfix · Dovecot · relay pair · DKIM/SPF/DMARC
Identity, directory & secrets
The cluster nobody lists and everybody depends on. Multi-master directory replication failure shows up as "users cannot log in" — usually on a Monday.
- Core practiceour practice
FreeIPA / Red Hat IdM cluster
FreeIPA · 389 Directory Server · replicas · CA renewal
- Core practice
OpenLDAP with syncrepl
OpenLDAP · syncrepl · delta replication
- Core practice
Single sign-on cluster
Keycloak · OIDC/SAML · HA across sites
- Core practice
Secrets management
HashiCorp Vault · Raft storage · auto-unseal · HSM
- Core practice
Service registry & mesh PKI
Consul · gossip encryption · ACL · Connect CA
- Core practice
Kubernetes secrets delivery
External Secrets Operator · Sealed Secrets · SOPS
- Regular delivery
Workload identity & SPIFFE
SPIRE · cert-manager · short-lived certs
- Regular delivery
Privileged access management
Teleport · bastion cluster · session recording
- Regular delivery
Zero-trust overlay access
WireGuard mesh · Tailscale · NetBird · policy
- Advisory
Active Directory domain controllers
AD DS · multi-site replication · FSMO · Azure AD Connect
Application & web clusters
The layer the business actually sees. Session handling and upgrade windows are where the downtime comes from.
- Core practiceour practice
LAMP / LEMP cluster
nginx + PHP-FPM + Galera · shared session store
- Core practice
Java application cluster
Tomcat · WildFly/JBoss EAP · JGroups · mod_cluster
- Core practice
CMS at scale
WordPress · Drupal · object-cache · read replicas
- Core practice
Session & state clustering
Redis-backed sessions · sticky routing · JWT
- Core practice
API backend cluster
FastAPI · Node · gRPC behind L7 LB · autoscaling
- Regular delivery
Shared / managed hosting platform
cPanel/WHM HA · Plesk · DirectAdmin · CageFS
- Regular delivery
Realtime media cluster
SRS · nginx-rtmp · LiveKit · Janus · TURN servers
- Regular delivery
Edge caching & CDN origin
Varnish · Cloudflare · Fastly · origin shielding
- Advisory
Jakarta EE application servers
WebLogic · WebSphere · cluster agents
- Advisory
IIS / .NET farm
IIS · ARR · shared config · session state
- Not ours
ERP platform administration
SAP Basis · NetWeaver · Solution Manager
CI/CD & artifact platforms
The pipeline is now production infrastructure. When runners are queued, every team stops at once.
- Core practiceour practice
CI runner fleet
GitLab Runner (autoscaling) · Jenkins agents · Buildkite
- Core practice
GitOps control plane
Argo CD · Flux · ApplicationSets · environment promotion
- Core practice
Artifact registry cluster
Harbor · Nexus · Artifactory · replication
- Core practice
Container registry HA
Harbor HA · distribution + object storage · GC policy
- Core practice
Build cache & shared storage
S3-backed cache · NFS cache · Bazel remote cache
- Regular delivery
Jenkins HA controller
Jenkins + Kubernetes agents · config as code
- Regular delivery
Progressive delivery
Argo Rollouts · Flagger · canary analysis
- Regular delivery
Supply chain security
cosign · SBOM · SLSA · Trivy in pipeline
Observability & monitoring
Long-term metric storage and alert routing are clustering problems, not dashboard problems.
- Core practiceour practice
Metrics cluster
Prometheus HA pairs · Thanos · Mimir · VictoriaMetrics
- Core practice
Logging cluster
Loki · Elasticsearch/OpenSearch · Fluent Bit · retention tiers
- Core practice
Dashboard & alerting HA
Grafana HA · Alertmanager cluster · silences
- Core practice
Kubernetes-native observability
Hubble · Cilium · kube-state-metrics · metrics-server
- Core practice
Synthetic & endpoint monitoring
Blackbox exporter · Uptime Kuma · probe nodes
- Core practice
Network & SNMP monitoring
LibreNMS · SNMP polling · flow analytics
- Regular delivery
Tracing cluster
Tempo · Jaeger · OpenTelemetry Collector
- Regular delivery
Legacy monitoring HA
Zabbix HA (server + proxies) · Icinga pair · Nagios
- Regular delivery
Cost observability
OpenCost · Kubecost · chargeback reports
- Regular delivery
On-call & escalation plumbing
Alertmanager routing · PagerDuty · escalation policy
Security operations clusters
Security tooling you cannot query under load is security tooling you will not use. Indexing and retention are capacity problems first.
- Core practice
Bastion & jump host cluster
Teleport · Apache Guacamole · session recording
- Core practice
Hardening baseline & compliance scan
OpenSCAP · CIS benchmark · Lynis · AIDE · auditd
- Regular deliveryour practice
SIEM cluster
Wazuh · Elastic Security · Splunk indexer cluster
- Regular delivery
IDS / IPS cluster
Suricata · Zeek · Snort · tap / mirror ports
- Regular delivery
Vulnerability scanning infrastructure
Nessus · OpenVAS · Trivy · scan scheduling
- Regular delivery
Network segmentation
Cilium network policy · microsegmentation · DMZ design
- Regular delivery
Endpoint & fleet inventory
osquery · Fleet · Ansible-driven compliance
- Advisory
Deception & honeypots
T-Pot · canary tokens
- Not ours
Managed detection & response staffing
24×7 SOC analyst coverage
Edge, SD-WAN & multi-site
Every site is a small cluster with a big dependency: the link. Design for the link being gone, not for it being fast.
- Core practice
WireGuard mesh / overlay networking
WireGuard · Tailscale · NetBird · ZeroTier
- Core practice
Multi-site L2/L3 extension
VXLAN · EVPN multi-site · stretched VLAN discipline
- Core practice
Edge Kubernetes fleet
K3s · k0s · KubeEdge · OpenYurt · offline-tolerant updates
- Core practice
Remote-site failover
local breakout · DC failover · link-state automation
- Core practice
Centralised fleet management
Rancher · Argo CD · SUSE Edge · GitOps at the edge
- Core practice
Site survivability
local cache · store-and-forward · degraded mode
- Regular deliveryour practice
SD-WAN overlay
VeloCloud · Cisco Catalyst SD-WAN · Fortinet Secure SD-WAN
- Partner-led
WAN optimisation & SD-Branch
vendor appliances · policy routing
- Partner-led
Private cellular & 5G core
CBRS · private 5G core · SIM provisioning
- Not ours
OT / ICS control systems
PLC programming · SCADA engineering · safety systems
Multi-cluster, migration & platform engineering
Where a fleet stops being "some clusters" and becomes a platform. This is the layer that determines your cost per workload.
- Core practiceour practice
Multi-cluster management
Karmada · Open Cluster Management · Rancher
- Core practice
Fleet provisioning
Cluster API · Terraform/OpenTofu · Ansible
- Core practice
Cloud → bare metal migration
cost modelling · workload porting · cutover plan
- Core practice
VMware → KVM / Proxmox migration
virt-v2v · wave planning · rollback rehearsal
- Core practice
Datacentre relocation
rack move · cluster evacuation · data replication
- Core practice
Infrastructure as code at fleet scale
Terraform · OpenTofu · Ansible · Terragrunt
- Core practice
Policy as code
Kyverno · OPA Gatekeeper · Conftest · CI gates
- Core practice
Capacity & cost engineering
utilisation analysis · right-sizing · headroom modelling
- Core practiceour practice
Cluster health assessment
38-point checklist · SPOF map · prioritised findings
- Regular delivery
Internal developer platform
Backstage · Port · golden paths · service catalogues
Found your cluster on this list?
Tell us which one is keeping you up at night. A thirty-minute call with an engineer, no deck, no obligation — and if it is one of the categories we do not own, we will say so on that first call.