monitor+server: Record million pixels processed metric

[yondonfu]

What does this pull request do? Explain your changes. (required)

Adds a million pixels processed metric. For now, this is only recorded on B by taking the total pixel count for a set of transcoded results, dividing it by 1m and then recording the result. This metric is helpful for monitoring pixel throughput on a B while also monitoring payment/ticket throughput since the latter should be based on the former.

The risk of overflow while using a float64 to record the metric should be minimal. Consider the following example:

• 179729280000 pixels per hour for a stream with 240p30fps, 360p30fps, 480p30fps, 720p30fps outputs (does not factor in decoded pixels from the input)
• 179729280000 / 1000000 = 179729.28 million pixels per hour
• 179729.28 million pixels per hour * 24 hours * 365 days = 1574428492.8 million pixels per year
• 1574428492.8 million pixels per year * 1000000 streams = 1574428492800000.0

So, with 1000000 streams running 24/7 for an entire year the million pixels value should still be below the max float64 value.

Specific updates (required)

• Add mil_pixels_processed metric in the monitor package
• Record mil_pixels_processed metric on B after counting the pixels in the transcoded results

How did you test each of these updates (required)

Manually.

Does this pull request close any open issues?

N/A

Checklist:

• Read the contribution guide
• make runs successfully
• All tests in ./test.sh pass
• README and other documentation updated
• Pending changelog updated

[darkdarkdragon]

LGTM