Release v0.0.114

Changes since v0.0.113:

* Certificates: add --json output flag to all `certificates get-*`
  commands. Emits structured JSON with separated cert (and key, where
  applicable) fields, replacing the awk-parsed PEM bundle. (#183)

* Azure: default the Azure cluster to minimal compute. (#181)

* Analytics proxy: proxy invalid Analytics URLs verbatim instead of
  rewriting them — disable nginx `merge_slashes` and drop the trailing
  slash from `proxy_pass` so empty path components are preserved.

Adjusted OpenShift istio gateway usage to be per-cluster.

Apply resources requests for OpenShift cluster node resources.

Refactor defaults for OpenShift clusters.

We now automatically created ingresses for OpenShift clusters
and the connect info automatically switches to using ingress
data for calculating connect info.  This is required as
OpenShift clusters are always in private subnets that arent
accessible anyways.

Adjust maximum kubernetes scaling time to 20 minutes.

In some cases, where the cluster itself needs to scale out and
add additional workers, this portion of the work can take 10
minutes on its own, separate from finally being able to
schedule and deploy the cluster on those new workers.

Fixed client traffic blocking to not hose clusters.

Switch back to using CAO 2.8.0 due to apparent bug.

Improve CAO deployment reliability by waiting for more potential cond…

…itions.

Update CAO default install to 2.9.0 during init.

Fixed agent creation for CAO clusters which use external network.