integration test fix 2

[leoluoInSea]

• local test working for TestJobCancellationWithoutSavepoint
• comment back in tests
• waits between cancel
• 5 min timeout
• additional logging
• remove local debug changes
• test app on flink 1.11
• remove local changes
• fix docker location
• move to proper dir
• break waiting after 30s so errors show
• log all events at failure
• forgot to add microk8s docker build and push
• wait longer for running app
• tm logs first
• see if microk8s can take the memory
• bump memory for tm due to 20mb allocated to task heap
• increase flink tm heap size
• set memory param. likely oom on second cluster
• fix typo
• change memory configs
• add limits to containers. add concurrency limit to go test
• limit cpu
• describe k8s resources on failure
• add get pods to view restarts
• try with minikube
• flink 1.8 not 1.11
• make executable
• remove microk8s from proxy kube
• fix minikube command
• disable rbac
• workaround auth
• fix file writing
• raise timeout
• all integ tests passing local direct
• bump minikube mem
• decrease minikube mem
• removing permissions crashes minikube api
• add role
• update timeouts
• raise timeout for cluster start. add tags to tests
• add better gc. add longer timeout
• comment out scale up test as it uses too much cpu
• try to increase cpus. 2 should be max but lets find out
• yep 2 cpus in max
• refactor clean up
• update readme
• update local dev docs
• update ubuntu
• update ubuntu
• see if kube config directory issue is due to ubuntu upgrade
• remove unused minikube command
• fix namespace in clusterrolebinding
• fix lint issues
• move import for lint
• upgrade ubuntu again
• attempt to fix ubuntu upgrade issue
• after cluster start
• check if kube dir exists
• set kube config location
• check kube config env var
• change the kubernetes config path