Validator metrics reporting

[ajhodges]

This PR enables support for our new metrics collection infrastructure based on the TICK stack. A validator registry (hosted at registry.network.shipchain.io) helps us track the state of every validator in our network, as well as manage contact info. This should help facilitate communication between ShipChain and validators. The metrics collected (system metrics, tendermint metrics, loom metrics) will be invaluable for monitoring network health as well as debugging issues among validators.

The changes to Hydra include:

• Rework of the hydra client set-info command.
  • Validates the presence of required parameters
  • Verifies that the current node is a validator
  • Also collect: email, primary contact name
  • Keeps a local cache of node 'info' so users don't have to re-enter it every time they want to update it
  • (If new user/validator) Register with the validator registry, create user if not already in the system
  • Update validator info in the registry (if user already exists, will prompt for registry password)
  • Fetch influxdb password for telegraf->influxdb authentication (metrics collection)
• Addition of hydra client enable-metrics command.
  • Sets hydra.validator_metrics = true setting
  • Installs telegraf
  • Configures rsyslog for system log collection (only collects shipchain-relevant logs)
  • Configures telegraf for rsyslog, prometheus inputs and influxdb output
• Addition of hydra client disable-metrics command.
  • Sets hydra.validator_metrics = false setting
  • Disables telegraf service
  • Stops rsyslog
• hydra client configure will now configure metrics if hydra.validator_metrics = true and hydra client set-info has been run
• Slightly rework setup.py requirements
• Version number bump

Note: metrics collection can always be opted out of by setting hydra.validator_metrics = false or running hydra client disable-metrics

[leewardbound]

This looks great, good flow on the registration, thanks