Improve e2e tests (#1111)

* icnrease vm size

* cache deps

* switch to the absolute cache path as cdp does not support shell expansion

* do not pull non-existing image

* manually install kind

* add alias to kind

* use full kind name

* one more name change

* install kind with other tools

* add bind mounts instead of copying files

* test fetching the runner image

* build image for pierone

* bump up the client-go version to match the master

* bump up go version

* install pinned version of kind before any test run

* do not overwrite local ./manifests during test run

* update the docs

* fix kind name

* update go.* files

* fix deps

* avoid unnecessary image upload

* properly install kind

* Change network to host to make it reachable within e2e runner. May not be the right solution though.

* Small changes. Also use entrypoint vs cmd.

* Bumping spilo. Load before test.

* undo incorrect merge from the master

Co-authored-by: Sergey Dudoladov <sergey.dudoladov@zalando.de>
Co-authored-by: Jan Mußler <janm81@gmail.com>

manifests/configmap.yaml

@@ -15,7 +15,7 @@ data:
   # connection_pooler_default_cpu_request: "500m"
   # connection_pooler_default_memory_limit: 100Mi
   # connection_pooler_default_memory_request: 100Mi
-  connection_pooler_image: "registry.opensource.zalan.do/acid/pgbouncer:master-9"
+  connection_pooler_image: "registry.opensource.zalan.do/acid/pgbouncer:master-11"
   # connection_pooler_max_db_connections: 60
   # connection_pooler_mode: "transaction"
   # connection_pooler_number_of_instances: 2
@@ -31,7 +31,7 @@ data:
   # default_memory_request: 100Mi
   # delete_annotation_date_key: delete-date
   # delete_annotation_name_key: delete-clustername
-  docker_image: registry.opensource.zalan.do/acid/spilo-12:1.6-p3
+  docker_image: registry.opensource.zalan.do/acid/spilo-12:1.6-p5
   # downscaler_annotations: "deployment-time,downscaler/*"
   # enable_admin_role_for_users: "true"
   # enable_crd_validation: "true"