DescribeInstanceGroup gets info about the specified instancegroup Exported so it can be used by the e2e tests, which don't want to instantiate a full cloudprovider.

Create a new CrossRequestRetryDelay.

ResizeInstanceGroup sets the size of the specificed instancegroup Exported so it can be used by the e2e tests, which don't want to instantiate a full cloudprovider.

Amazon recommends having no more that 40 volumes attached to an instance, and at least one of those is for the system root volume.

Default volume type for newly created Volumes TODO: Remove when user/admin can configure volume types and thus we don't need hardcoded defaults.

We sometimes read to see if something exists; then try to create it if we didn't find it This can fail once in a consistent system if done in parallel In an eventually consistent system, it could fail unboundedly MaxReadThenCreateRetries sets the maximum number of attempts we will make.

Service annotation specifying the protocol spoken by the backend (pod) behind a secure listener.

Service annotation requesting a secure listener.

Annotation used on the service to indicate that we want an internal ELB.

Annotation used on the service to enable the proxy protocol on an ELB.

Service annotation specifying a comma-separated list of ports that will use SSL/HTTPS listeners.

The tag name we use to differentiate multiple logically independent clusters running in the same AZ.

The tag name we use to differentiate multiple services.

The tag name used on a subnet to designate that it should be used for internal ELBs.

The tag name used on a subnet to designate that it should be used for internet ELBs.

AWSCloud is an implementation of Interface, LoadBalancer and Instances for Amazon Web Services.

Backoff manages a backoff that varies based on the recently observed failures.

CrossRequestRetryDelay inserts delays before AWS calls, when we are observing RequestLimitExceeded errors Note that we share a CrossRequestRetryDelay across multiple AWS requests; this is a process-wide back-off, whereas the aws-sdk-go implements a per-request exponential backoff/retry.

This is a simple pass-through of the Autoscaling client interface, which allows for testing.

Abstraction over AWS, to allow mocking/other implementations.

TODO: Should we rename this to AWS (EBS & ELB are not technically part of EC2) Abstraction over EC2, to allow mocking/other implementations Note that the DescribeX functions return a list, so callers don't need to deal with paging.

Abstraction over the AWS metadata service.

This is a simple pass-through of the ELB client interface, which allows for testing.

InstanceGroupInfo is returned by InstanceGroups.Describe, and exposes information about the group.

InstanceGroups is an interface for managing cloud-managed instance groups / autoscaling instance groups TODO: Allow other clouds to implement this.

Volumes is an interface for managing cloud-provisioned volumes TODO: Allow other clouds to implement this.