Bitpoke Mysql Operator Versions Save

• Add support for pt-kill to limit queries
• Fixing issue with pod labeling
• Serve backups using http instead of ncat and use basic auth
• Add owner reference to PVC to be deleted when a cluster is deleted
• Add support for PodDisruptionBudget by adding minAvailable cluster param
• Set operator to listen by default on all namespaces
• Set a limit for how many recurrent backups to keep by specifying backupScheduleJobsHistoryLimit cluster param

• .Status.BackupUri is updated now, and .Spec.BackupUri is not modified
• Bump orchestrator chart version 0.1.3
• Set Pod labels according to Orc status
• Remove work items from queues on cluster deletion
• Recurrent backups are done by operator using a cron library and not by using k8s cronjob

• Add master endpoint foo-mysql-master that points to master node.
• Auto acknowledge an audit recovery in orchestrator.
• Add openapi validation to CRDs.
• Add pt-heartbeat for every node and cluster lag monitoring.
• Add nodes conditions from orchestrator in cluster status.
• In foo-mysql service register just healthy nodes.
• Create a service with all nodes foo-mysql-nodes (the new headless service for statefulset).
• Add maxSlaveLatency to cluster spec, that represents the allowed latency for a node.
• Rename crd apiVersion from mysql.presslabs.net to mysql.presslabs.org
• Add --install-crds flag to operator and update chart to use installCRDs.

• bug fixing

• add support for prometheus-operator, by exposing scraping ports to the service

• allow running controller in different namespace from cluster
• orchestrator topology credentials are passed from controller to cluster, rather than mounting the secret into cluster pods

Initial operator release