Getting the following error in the logs.
`{"level":"error","msg":"agones.dev/agones/pkg/client/informers/externalversions/factory.go:74: Failed to watch *v1alpha1.GameServer: unknown (get gameservers.stable.agones.dev)","time":"2018-07-19T21:52:23Z"}`

Needed the `watch` permission to make it go away.

We aren't watching for events right now, so it's just noise,
but it's better with it gone.

Includes some tweaks to the release process as well.

Extended the gRPC SDK to include a `GameServer()` function for both Go, C++
and REST by extending the backing `sdk.proto` and implementing the functionality
in the sdk sidecar.

This is the start of work needed for #277.

Added information about project: https://github.com/dgkanatsios/AksNodePublicIPController

This allows the `GameServer` (and anything that uses it as a template,  such
as `Fleets`), to provide configuration and assignment of multiple ports to the
game server container.

For example:

```yaml
apiVersion: "stable.agones.dev/v1alpha1"
kind: GameServer
metadata:
  name: "simple-udp"
spec:
  ports:
  - name: default
    portPolicy: "dynamic"
    containerPort: 7654
  template:
    spec:
      containers:
      - name: simple-udp
        image: gcr.io/agones-images/udp-server:0.1
```

This also includes update to documentation - but in a separate bottom section
of each document, to allow for moving into the live version at release time.
(This is becoming a mess - a website can't come soon enough).

This also adjusts the `GameServer > Status`, to be able to handle multiple
ports. It now looks like this:

```
Status:
  Address:    192.168.99.100
  Node Name:  agones
  Ports:
    Name:  default
    Port:  7502
  State:   Ready
```

While this a breaking change (due to the change in `GameServer > status`, the
previous legacy configuration of still works (even thought it's no longer
documented), as it is automatically converted to the new version.

Plans are to remove this legacy conversion functionality in 0.4.0, since being
in alpha, we have a very low commitment to backward compatibility.

As an aside - versioning for CRDs is coming soon, so this will also solve this
type of problem long term.

It looks like the `foregroundDeletion` finalizer that gets added to the
GameServer when a deletion with foreground propogation is requested -
when it gets removed, by the Kubernetes system, the ResourceVersion/Generation
doesn't get incremented.

This means when we do an `Update` it will go through (usually it fails if the
ResourceVersion/Generation has changed since the last update), rather than
failing and going back to re-sync.

This can cause a race condition where the `Update` can basically put back
the `foregroundDeletion` finalizer, if it gets removed between retrieving the
`GameServer` object and doing the `Update()` operation.

We actually don't need the `foregroundDeletion` propogation on `Delete()`, as
we only want to keep the `GameServer` around until the backing `Pod` is gone,
so we can actually just turn this off instead in the code.

As a note - also explored using a `Patch()` operation to try and solve this
problem. AFAIK, a patch (and specifically a JSONPatch - other patch types
didn't seem better), can only do index based removal from a list of items.

This could just present other types of race conditions, as finalizers could
be added/removed by other systems at the same time as well, changing the index
to the finalizer we want at the same time.

Not sure if this is a bug in the Kubernetes system, or working as intended,
but this solves the issue.

Update codegen library to 1.10.5, and regen the CRD clients.

This may be the cause of #278 (but that change needed to happen anyway)

Keeping up with Kubernetes versions, as well as providing
us the functionality to do SubResources.

Also will enable tightening up of permissions re: #150

Found a line here that makes no sense.

Have working instructions for running Kubernetes 1.10 on top of
minikube 0.28.0 - including a workaround for non default profiles.

Since Agones works fine on k8s 1.10, figured it was worth making the move.

Implemented workerqueue for node modifications in the
PortAllocator, so that if the master happens to go
down, then the operations the PortAllocator has to do
to keep things in check will retry and result in a
correct state.

Closes #60

- (Conflict merge) Update Helm App Version
- update doc as requested

Implement grpc-gateway in front of the gRPC based sdk-server, so that
it can be access via HTTP+JSON.

This includes documentation and a swagger/openapi specificiation.

This also has been implemented such that the sdk-server is still a single
binary, and as such, the HTTP+JSON interface can still be used for local
development.

Closes #240

Some more tutorials to help people get started.

Missed that this should be 0.3.0

This should hopefully fix these tests.

There was a race condition in the original
implementation. This fixes that issue!

(and makes the code cleaner!)

Trying to stop timeouts :/

This also replaces the C++ build image we have from the gRPC docker hub,
with one that we manage ourselves, since getting updates in a timely
matter was getting difficult.

The C++ base will likely go away once the transition to REST based SDK,
so the base may end up going away.

This is prerequisite for #240

Just keeping up with updates.

I think this will stop this test being flakey.

I believe this shoud fix this flakey test

Realised that all the new Rust SDK code
didn't have the Apache licence applied.

This applies the licence.

Make it clearer to download the release archive
from releases, rather than an implied `git checkout`

Closes #248

- Include Rust in the list of SDKs
- Also put a development warning on the SDK

This should make Cloud Builder faster
- Request an 8 core machine
- Move the test and build steps to be parralel.

Noticed that this library wasn't locked to a version,
so when I attempt to regen the code, I got some incompatible
code.

This fixes things.