diff --git a/README.md b/README.md index 0ea0dbf78..6538d39fa 100644 --- a/README.md +++ b/README.md @@ -31,26 +31,19 @@ state and over time move to full "production readiness". The following table has the general readiness state, the table below covers individual newer features separately. -| Platform | Kubernetes version | ArangoDB version | ArangoDB K8s Operator Version | State | Production ready | Remarks | -|----------------------|--------------------|------------------|-------------------------------|-------|------------------|-----------------------| -| Google GKE | 1.12 | >= 3.3.13 | | Runs | Yes | Don't use micro nodes | -| Google GKE | 1.13 | >= 3.3.13 | | Runs | Yes | Don't use micro nodes | -| Amazon EKS | 1.11 | >= 3.3.13 | | Runs | Yes | | -| Pivotal PKS | 1.11 | >= 3.3.13 | | Runs | Yes | | -| IBM Cloud | 1.11 | >= 3.4.5 | >= 0.3.11 | Runs | Yes | | -| IBM Cloud | 1.12 | >= 3.4.5 | >= 0.3.11 | Runs | Yes | | -| IBM Cloud | 1.13 | >= 3.4.6.1 | >= 0.3.11 | Runs | Yes | | -| Amazon & Kops | 1.10 | >= 3.3.13 | | Runs | No | | -| Azure AKS | 1.10 | >= 3.3.13 | | Runs | No | | -| OpenShift | 1.10 | >= 3.3.13 | | Runs | No | | -| Bare metal (kubeadm) | 1.10 | >= 3.3.13 | | Runs | Yes | | -| Bare metal (kubeadm) | 1.11 | >= 3.3.13 | | Runs | Yes | | -| Bare metal (kubeadm) | 1.12 | >= 3.3.13 | | Runs | In progress | | -| Bare metal (kubeadm) | 1.13 | >= 3.3.13 | | Runs | Yes | | -| Bare metal (kubeadm) | 1.14 | >= 3.3.13 | | Runs | In progress | | -| Minikube | 1.10 | >= 3.3.13 | | Runs | Not intended | | -| Docker for Mac Edge | 1.10 | >= 3.3.13 | | Runs | Not intended | | -| Scaleway Kubernetes | 1.10 | >= 3.3.13 | ? | No | | | +| Platform | Kubernetes Version | ArangoDB Version | ArangoDB Operator Version | State | Remarks | Provider Remarks | +|---------------------|--------------------|------------------|---------------------------|-------------|-----------------------|------------------------------------| +| Google GKE | 1.14 | >= 3.3.13 | | Production | Don't use micro nodes | | +| Google GKE | 1.15 | >= 3.3.13 | | Production | Don't use micro nodes | | +| Azure AKS | 1.14 | >= 3.3.13 | | Production | | | +| Azure AKS | 1.15 | >= 3.3.13 | | Production | | | +| Amazon EKS | 1.14 | >= 3.3.13 | | Production | | [Amazon EKS](./docs/providers/eks) | +| IBM Cloud | 1.14 | >= 3.4.6.1 | >= 0.3.11 | Production | | | +| OpenShift | 3.11 | >= 3.3.13 | | Production | | | +| OpenShift | 4.2 | >= 3.3.13 | | In Progress | | | +| BareMetal (kubeadm) | 1.14 | >= 3.3.13 | | Production | | | +| Minikube | 1.14 | >= 3.3.13 | | Devel Only | | | +| Other | 1.14 | >= 3.3.13 | | Devel Only | | | Feature-wise production readiness table: diff --git a/docs/README.md b/docs/README.md index bff3a892a..b3c4b5d2c 100644 --- a/docs/README.md +++ b/docs/README.md @@ -3,3 +3,4 @@ - [Tutorial](https://www.arangodb.com/docs/stable/tutorials-kubernetes.html) - [Documentation](https://www.arangodb.com/docs/stable/deployment-kubernetes.html) - [Design documents](./design/README.md) +- [Providers](./providers/README.md) diff --git a/docs/providers/README.md b/docs/providers/README.md new file mode 100644 index 000000000..d59b40570 --- /dev/null +++ b/docs/providers/README.md @@ -0,0 +1,3 @@ +# Supported Providers + +- [Amazon EKS](./eks/README.md) \ No newline at end of file diff --git a/docs/providers/eks/README.md b/docs/providers/eks/README.md new file mode 100644 index 000000000..2a856a4fc --- /dev/null +++ b/docs/providers/eks/README.md @@ -0,0 +1,26 @@ +# Amazon AWS Remarks + +## Elastic Block Storage + +Documentation: +- [AWS EBS Volume Types](https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/ebs-volume-types.html) + +Remarks: +- It is recommended to use at least GP2 (can be IO1) volume type for ArangoDeployment PV. +- GP2 Volume IOPS is mostly based on storage size. If bigger load is expected use bigger volumes. +- GP2 Volume supports burst mode. In case load in ArangoDeployment is expected only periodically you can use +smaller GP2 Volumes to save costs. +- AWS EBS support resizing of Volume. Volume size can be changed during lifetime, but it requires pod to be recreated. + +## LoadBalancer + +Documentation: +- [AWS LB Annotations](https://kubernetes.io/docs/concepts/cluster-administration/cloud-providers/#load-balancers) + +Remarks: +- AWS LB in TCP mode is able to resend request in case of timeout while waiting for response from Coordinator/DBServer. +This can break some POST requests, like data insertion. To change default value, set to 60s, +you can set annotation for ArangoDeployment LoadBalancer service. +``` +kubectl annotate --overwrite service/-ea service.beta.kubernetes.io/aws-load-balancer-connection-idle-timeout= +``` \ No newline at end of file