Google Opens Patches For "METRICFS" That They Have Used Since 2012 For Telemetry Data

[phoronix]

Phoronix: Google Opens Patches For "METRICFS" That They Have Used Since 2012 For Telemetry Data

The METRICFS file-system has been in use internally at Google since 2012 for exporting system statistics to their telemetry systems with around 200 statistics being exported per machine. They are now posting the METRICFS patches as open-source for review and possible upstreaming...

http://www.phoronix.com/scan.php?page=news_item&px=Google-RFC-Open-Source-METRICFS

[bachchain]

[andrei_me]

I.e: Google wants to upstream their private work, trying to shoehorn it as a standard because it exists for longer, reduce their own burden to maintain it instead of allowing an open from beginning code from RH becoming the standard and Google would have to deal with their legacy 😒

[eydee]

Canonical: Heavy breathing.

[ElectricPrism]

So Google made this thing 8 years ago and now they suddenly want to "open it up to the community",

What, does the advantage of leaving it in-house close no longer exist? -- and now the advantage is to reduce maintenance costs or prevent a competitor from standardizing a new system that differs from theirs?

These days with Chrome tracking you up the wazoo, Google in bed with China there's not much they are doing that is interesting or trustworthy anymore, especially after creating Echo chambers and soaking up advertising money while trying to sway political elections to reinforce their own monopoly and continue to dismantle and consume competitors as seen in their Congress hearings with other Tech Giants sweating like they were in the sauna while being grilled alive.

[yoshi314]

hm,, this might make monitoring of linux machines even easier. i wonder what sorts of metrics are exposed and can they be extended to cover e.g. containers.

[polarathene]

hm,, this might make monitoring of linux machines even easier. i wonder what sorts of metrics are exposed and can they be extended to cover e.g. containers.

Doesn't Netdata and Prometheus cover that pretty well already? Both are quite capable of gathering system metrics and doing so with low overhead afaik. For containers, there's googles cadvisor (https://github.com/google/cadvisor), which you can collect stats from to Prometheus .

[yoshi314]

Doesn't Netdata and Prometheus cover that pretty well already? Both are quite capable of gathering system metrics and doing so with low overhead afaik. For containers, there's googles cadvisor (https://github.com/google/cadvisor), which you can collect stats from to Prometheus .

yes, but prometheus pulls data from exposed metric endpoints. and this might offer a more lightweight way to get at them. the less middlemen the better.

[CochainComplex]

two of the big ones....do you offer beer aswell?

[schmidtbag]

In some ways I'm thinking "oh that's nice, maybe now that we know how this all works, we can take off our tin-foil hats" but then I realized "Google probably isn't using the same code internally".

Not that I care anyway. Google can go ahead and collect my data, as long as it doesn't make my life worse.