router-netsender: report statistics from GLMIFI

[Trek Hopton]

Currently the router-netsender reports system logs to the cloud.

In order to gain more insights into the operation and potential issues of the GLMIFI, it would be useful to have access remotely to some of the operational statistics.

Of particular interest are:

• upload/download
• ping
• CPU usage
• Memory usage
• Uptime

As of writing this issue a number of different strategies have been explored, detailed here.

The currently proposed approach is to report stats using netsender pins so that we can store and view the data with our currently existing cloud tools eg. graph over time using vidgrind’s search page.

This issue should be resolved when the statistics listed above have been explored and either; implemented, or ruled out as impractical to implement.

Individual issues will be created for each statistic.

‌

[Trek Hopton]

Upon further thought, and considering the urgency of this feature for our remote Whyalla deployment, it has been decided that the best approach for now is to run relevant commands via SSH and send the output as text to the cloud. There it can be parsed and analysed.

The shell commands give inconsistent and somewhat complex outputs so the ability to change the parsing functions after deployment is important.

The shell commands that will be used are ping, uptime, and top, which will provide us with text output containing ping, uptime, CPU and memory usage.

A somewhat separate feature is to add upload/download monitoring to this netsender. This can be addressed in another issue.

My approach to collecting the command output will be as follows: get router shell command output using existing ssh and remote packages > create a map of command name string to output string > encode map as JSON and send on pin T3.

The handling of this JSON data can be addressed in another issue.

[Trek Hopton]

Resolved in PR #92.