A lot of the changes in apiserver could have been represented more
cleanly - this returns the signatures to their older behavior (and
unbreaks OpenShift).
A namespaced resource that supports ALL may allow creation
on the root (all namespaces) collection, thus adding POST
here.
We need to better formalize the definition of calls on namespaced
resources at the root scope, so Storage objects that do not support
that call pattern can do so at definition time and reject those
calls.
Connecter is a type of resource that connects a request
coming from the client to an internal request within the cluster.
It will be used for exposing a pod's proxy, exec, and portforward
endpoints.
Allows REST consumers to build paths like:
/api/v1beta3/namespaces/foo/webhookresource/<name>/<encodedsecretinurl>
Also fixes parameter exposure for subresources (was only fixed for
v1beta3).
In addition to Getter interface, API Installer now supports a
GetterWithOptions interface that takes an additional options object when
getting a resource. A flag is now returned from rest.ResourceStreamer
that indicates whether the streamed response should be
flushed when written back to the client. This is to support log
streaming.
Dependency chain is now api -> api/rest -> apiserver. Makes the
interfaces much cleaner to read, and cleans up some inconsistenties
that crept in along the way.
This commit adds support to core resources to enable deferred deletion
of resources. Clients may optionally specify a time period after which
resources must be deleted via an object sent with their DELETE. That
object may define an optional grace period in seconds, or allow the
default "preferred" value for a resource to be used. Once the object
is marked as pending deletion, the deletionTimestamp field will be set
and an etcd TTL will be in place.
Clients should assume resources that have deletionTimestamp set will
be deleted at some point in the future. Other changes will come later
to enable graceful deletion on a per resource basis.
PUT /api/v1beta3/namespaces/default/pods/foo/status
{
"metadata": {...}, // allowed for valid values
"spec": {}, // ignored
"status": {...}, // allowed, except for Host
}
Exposes the simplest possibly change. Needs a slight refactoring
to RESTUpdateStrategy to split merging which can be done in a
follow up.
prometheus library for monitoring, which exports some basic resource
usage metrics by default, like number of goroutines, open file
descriptors, resident and virtual memory, etc. I've also started adding
in request counters and latency histograms, but have only added them to two
of our HTTP handlers. If this looks reasonable, I'll add them to the rest
in a second PR.
Add a route for /proxy/<kind>/<name>/ without requiring additional path
information. This allows for e.g. /proxy/service/myservice/, which will
proxy to the service at /, instead of always requiring something in the
path.