How to build dashboard for single prometheus instance have multiple hosts?

I’m trying to make a dashboard that visualizes swarm metrics that scrape using remote Prometheus node.I have 4 swarm nodes running cadvisor and node-exporter as global service ,exposed via nginx on port 80 (Loadbalancer) .

My promthues config

  - job_name: 'stage_swarm'
    scrape_interval: 60s
    scrape_timeout: 59s
    metrics_path: /metrics
    scheme: http
    - targets:
      - cadvisor.stage:80
      - exporter.stage:80 (  metrics will be collected from  remote swarm cluster node-exproter,via nginx(LB)  )

I could see that node_meta has all the swarm nodes listed ,but am not able to visualize all the nodes separately .

How do I create a dashboard for all the host inside node_meta ? When I query using node_name or job ,I always get node count as 1 .

count(count(container_tasks_state{job="stage_swarm"}) by (container_label_com_docker_swarm_node_id)) – always  1
sum((node_memory_MemTotal - node_memory_MemFree - node_memory_Cached - node_memory_Buffers - node_memory_Slab) * on(instance) group_left(node_name) node_meta{job="stage_swarm"}) by (node_name)
sum(node_memory_MemTotal * on(instance) group_left(node_name) node_meta{job="stage_swarm"}) – always wrong. Same issue for all metrics.