Manual

Good points:

• Old school approach
• Manually moving the files to the back end server
• Loading files at say the end of the day

Bad points:

• Manually intensive
• No chance of graphs that are live current graphs
• Graphs only available after the file updates

Automate file transfers

Good points:

• Semi-automated approach
• Use common UNIX/Linux tools to move the data files
• Loading files at say the end of the day or periodically

Bad points:

• No chance of graphs that are live current graphs
• Graphs only available after the file updates

Fully working architecture

- up to the minute graphs

- with security

Good points:

• Fully automatic
• Install njmon, ssh keys and a cron job
• Secure - but performance data is not sensitive
• Use network sockets to feed the data into the injector

Bad points:

• You have to set up ssh between Back end and each Endpoint

Fully working architecture

- up to the minute graphs

- with simpler setup

Good Points:

• Fully automatic
• Install njmon and a cron job
• need to run the njmon_collector on the back end
• Not so secure as ssh mode(initial connect is encrypted lightly but the data is in the clear)
• Use network sockets to feed the data into the collector
• Optionally, run the injector or save to a JSON file or both

Bad points:

• You have to set up ssh between Back end and each Endpoint