|
|
|
| Link |
https://aws\.amazon\.com/eks/ |
https://cloud\.google\.com/kubernetes\-engine |
|
|
|
|
|
|
| 1.18 |
✖︎ |
✖︎ |
| 1.17 |
✖︎ |
Preview |
| 1.16 |
✔️ |
✔️ |
| 1.15 |
✔️ |
✔️ |
| 1.14 |
✔️ |
✔️ |
| 1.13 |
|
|
| Notes |
|
|
| Notes |
Notes |
|
|
|
|
|
|
|
| Max number of clusters per region |
100 |
50/zone + 50 regional clusters |
| Max nodes per cluster |
3000 |
5000 |
| Max nodes per node pool |
100 |
1000 |
| Max node pools |
30 |
No documented |
| Max pods per Node |
737 |
110 |
| Notes |
|
|
| Notes |
Notes |
|
|
|
|
|
|
|
| Control plane |
10 cents per hour per control plane |
10 cents per hour per control plane |
| Notes |
|
|
| Notes |
Notes |
|
|
|
|
|
|
|
| Control plane upgrades |
Manual |
Automatic + Manual |
| Worker nodes upgrades |
Manual |
Automatic + Manual |
| Notes |
|
|
| Notes |
Notes |
|
|
|
|
|
|
|
| Operating system |
Amazon Linux, Ubuntu, Amazon Machine Image (AMI), Windows server and Bring-Your-Own-OS, Bottlerocket OS |
Container Optimised OS, Ubuntu, Windows Server |
| Container runtime |
Docker |
Docker, containerd |
| Container Networking |
Amazon VPC CNI |
Native GKE CNI |
| Sandbox |
Not available |
gVisor |
| Bare metal nodes support |
✔️ |
✖︎ |
| GPU nodes |
✔️ |
✔️ |
| TPU nodes |
✖︎ |
✔️ |
| Cluster Autoscaling |
✔️ |
✔️ |
| Node Auto-repair |
✖︎ |
✔️ |
| Notes |
|
|
| Notes |
Notes |
|
|
|
|
|
|
|
| Secrets |
Encrypted at rest with AWS KMS |
Encrypted at rest with Cloud KMS |
| Key for encryption |
Configurable |
Configurable |
| Network policy support |
Yes (Calico) |
Yes (Calico) |
| Notes |
|
|
| Notes |
Notes |
|
|
|
|
|
|
|
| SLAs |
99.95% |
99.5% (zonal), 99.95% (regional) |
| Notes |
|
|
| Notes |
Notes |
|
|
|
|
|
|
|
| Terraform support |
✔️ |
✔️ |
| Notes |
|
|
| Notes |
Notes |
|
|
|
|
|
|
|
| Compliance |
HIPAA, SOC, ISO, PCI DSS |
HIPAA, SOC, ISO, PCI DSS |
| Certified Kubernetes |
✔️ |
✔️ |
| Notes |
|
|
| Notes |
Notes |
|
|
|
|
|
|
|
|
|
|
|
|
|
| Vertical Pod Autoscaling |
✔️ |
✔️ |
| Master/control plane high availability options |
Control plane is deployed across multiple Availability Zones |
Zonal/multi-zonal clusters: single control planeRegional clusters: control plane replicas in 3 zones |
| Cost Dashboard per Namespace/Cluster |
|
✔️ |
| Master Authorized Network |
Public + Private Endpoint |
Public + Private Endpoint |
| Resource monitoring |
No (Deploy Prometheus manually) |
Cloud Monitoring (Former Stackdriver) |