Fix major performance issues with ruby_gyr_reader_tier

[truedat101]

ruby_gyr_reader_tier suffers from some performance challenges:

• It takes about 99% of CPU core 1 when running
• It may spend a bunch of that time doing debug output logging
• It may spend time simply busy blocking
• It may be doing too much work in the reader tier
• It may need to utilize threads to avoid hogging too much CPU
• It may not be reading enough tags quickly enough to seem responsive as a webapp (time between read to actually logging in the web UI seems sort of laggy), should be less than 1 second to update

[truedat101]

@davicitoafc I think the very first thing to do is disable all of the excess logging that goes into each compute cycle. Right now we have a processing loop that occurs with each reader scan. At that point, we do a bunch of steps. All of those steps have all of the debug output turned on . Turn off (1) web server debug (2) database ORM debug (3) comment out any print statements we've added. I would suggest documenting how to turn the debug stuff on and off so it's easy to get it back. But I think for this round of effort, let's see if that reduces the time spent in each processing loop (so we don't miss reads).

[davicitoafc]

After removing debug I only notice a slight difference in lag @truedat101.

[truedat101]

Can you ensure there is a way to turn debug on-off easily? That was the request. Clearly you've benefited from having the debugging.

My request is:

• make it easy to turn debugging on/off from the command line. Since you are a ruby person, I'd suggest either use log4r, or just use the builtin logger: http://ruby-doc.org/stdlib-2.6.3/libdoc/logger/rdoc/
• The same way we use graylog on our server stuff on the other project we work on, the configuration allows you to turn on/off logging without using comment to remove loggin.
• document the way to activate logging ... I typically use info, debug, error, trace.

[davicitoafc]

@truedat101 Not sure what you mean by turn "off" or "on"

[truedat101]

So most logging systems have a way to deactivate them so you never see anything logged. To turn off, usually one would change a configuration file or pass in a debug flag as an argument, or set an environment variable.

Please review this link so you can understand the basics of using a logger and logger priorities: https://www.thoughtco.com/write-log-messages-in-ruby-2908323

The example they have there shows the log level being set to WARN. That means anything of lower priority isn't shown. That's how you'd turn off DEBUG level meesages.

My suggestion is replace every single puts() call with the logger calls. And then put log level in the yaml file.