Metrics renamed, dashboad example broken

[bbeattie-phxlabs]

In 2018, node exporter renamed some metrics (see https://github.com/prometheus/node_exporter/releases/tag/v0.16.0, https://www.robustperception.io/new-features-in-node-exporter-0-16-0). As a result, a number of dashboard metrics are broken such as https://github.com/prometheus-community/elasticsearch_exporter/blob/master/examples/grafana/dashboard.json#L1419 which refers to a now absent node_memory_MemTotal instead of node_memory_MemTotal_bytes.

[sysadmind]

I wonder if these metrics are useful as defaults for the dashboard. There are a lot of ways to run elasticsearch and many of them do not map directly to an individual host/node (ex Amazon OpenSearch, Elastic Cloud, Docker). I think the best option may be to remove them from the default dashboard and allow users to customize the dashboard to their environment.

[bbeattie-phxlabs]

Leave the Elasticsearch graphs specific to Elasticsearch data; Makes sense to me.

[QuanTran91]

I face the same issue, should we replace node_memory_MemTotal by node_memory_MemTotal_bytes in the default dashboard?