Rework the TIG server role

[marcus-grant]

The TIG stack role(s) shouldn't be a single role, that only makes sense for singular servers with no cross-server monitoring.

• There should be one or a few servers that take a server role that collects, stores, and displays the monitoring data in dashboards
  • This will involve an influxDB set of tasks, or possibly its own role (probably not)
  • A grafana server setup which sets up the grafana container and copies over config files to setup its security, dashboard configurations, listeners, etc.
• There should be a role for setting up multiple types of telegraf plugins and agents on multiple machines that will send data to the one or few monitoring servers
  • Possibly there should be multiple telegraf roles and containers for each monitoring plugin and configuration of interest:
  • RSyslog
  • Temperatures
  • Hardware Loads
  • Docker monitoring
  • HTOP like monitoring of threads with their cpu time, state, and memory
  • Network activities like current and accumulated bandwidth

[marcus-grant]

Not pursuing this anymore, prometheus is the stack I'll use in the future.