Summary: 26 instances, 26 unique Text Count # TODO we some k8s template still using the 'dashboard_host' 1 // TODO: hardcode here, change it when has better solution 1 host = alert_manager_hosts[0] # TODO not sure if alert manager with HA workds this way 1 // TODO: Insert code to detect warnings. 1 # TODO on premise, using ip as "nodename" 1 # TODO nodename == hostname on aks 1 // TODO: Only Restart Service instead of exit whole process and Restart by external system. 1 // TODO: Store AttemptId in AMStatus, and double check it before pushStatus 1 // TODO: make it a cluster-wise config 1 //TODO: Node Gpu policy filter the nodes; 1 # TODO: Change the command with linux_shell.execute_shell to docker lib. 1 //TODO: apply other node selection policy in the future; 1 # TODO: tell openpai / HiveD to keep idle when scaling?? 1 // TODO: Only Restart WebServer instead of exit whole process and Restart by external system. 1 // TODO: workaround for circular dependencies, need redesign module structure 1 // TODO: align same format of jobname with each submit ways 1 // TODO: Implement Service Rolling Upgrade 1 cmd_timeout = 10 # TODO 99th latency is xxx 1 // TODO: update grouplist at initialization 1 // TODO: replace updateGroup2ExnternalMapper 1 // TODO: Update TaskStatus.ContainerIsDecommissioning 1 # TODO: this piece of code seems not corret, gpu_mem_util is 1 # TODO speed this up, since this is O(n^2) 1 // TODO: use other policy to update index 1 # TODO check remot links 1 # TODO: + azure information 1