Add support for MPPT metrics

[sveba]

Summary

• Adds MPPT (Maximum Power Point Tracking) metrics. MPPT metrics are archived by the fronius inverter every 5 mins.
• This results in an additional API request to the Fronius device with the same --symo.timeout
• Each API request is done sequentially. That means the total scrape time can be 2x --symo.timeout. Please take that into consideration when configuring Prometheus scrape interval
• Adds 2 new flags: --symo.enable-power-flow, --symo.enable-archive with default values being true. With these flags, you can also disable each API endpoint.

Breaking Changes

• The --symo.url flag suggested to use the full URL path like --symo.url http://symo.ip.or.hostname/solar_api/v1/GetPowerFlowRealtimeData.fcgi previously. This has now changed to base host name only, e.g. --symo.url http://symo.ip.or.hostname (without trailing slash)

Checklist

• Categorize the PR by setting a good title and adding one of the labels:
  fix, enhancement, documentation, change, breaking, dependency
  as they show up in the changelog
• Update documentation.
• Update tests.

[ccremer]

Hi!
Thanks for the PR. I'll try to review it today or in the weekend.

[ccremer]

Good work so far! 🎉
I think it could be even more improved for both user and maintainers though.

[ccremer]

2 things that came to mind: Right now the timeout option is applied to both requests. Meaning a timeout of 5 seconds could potentially end up in a scrape of 10 seconds. If we have further Data to scrape, 15 seconds etc. It should be at least documented what this means.

However, thinking even further, it's probably a good idea to make the requests in parallel. That should end up in faster scraping. A little more care needs to be given regarding error handling, in order to present only 1 error to the user, I'd propose to add https://github.com/hashicorp/go-multierror into the mix: Each fatal scrape error can be combined into one and it nicely displays multiple occurred errors.

I would be fine with doing this in a second step in a separate PR. If you would be interested in doing this, I would also delay a release until we have parallelization.

[sveba]

I would be fine with doing this in a second step in a separate PR. If you would be interested in doing this, I would also delay a release until we have parallelization.

Agree, but additional PR would be my choice

[ccremer]

Thanks for the work!
Do you know if you are going to work on the parallelization right away or would you like a prerelease first?

[sveba]

Thanks for the work!
Do you know if you are going to work on the parallelization right away or would you like a prerelease first?

I'm leaving in vacation for 1,5 months so I would not be able to work on this for that time. After that no problem.
About release: like I said, this is your baby, you can decide. I have the exporter running on my server so I'm not in a hurry. :)

[ccremer]

Ah ok. I'll wait for a release and try to come up with something, shouldn't be too hard.
Enjoy your vacation and thanks again for your contribution!

[sveba]

Na klar. Habe ich gerne gemacht. Wie gesagt, bin absoluter Go-Anfänger ;)