That’s a weird emoji to use for elixir

GIGO

…I feel like openssh has a much larger attack surface than a simple binary.

If you’re going to this extent already, you may as well jump on the run0 approach systemd is introducing.

oh no, I can hear rumbling

You don’t need to be skilled in something to get enjoyment out of it

Likewise you don’t necessarily get enjoyment out of something just because you’re skilled in it.

void

Does that not put JS (node) back on the table?

I’d say it’s the low level language doing the heavy lifting, python or JS in this scenario are just front-ends.

Hell, I think FORTH has C bindings, that’s not power, that’s mental illness

IIRC JavaScript + TypeScript is the biggest demographic of engineers in the industry if you go by GitHub stats

I suppose you could call that power in a way

Never go full APL

releases a change where all routes accidentally go to the error page controller

🤷‍♂️ They’re 4xx errors, won’t be us

Well I didn’t say anything about perfectly clean, but I agree, it’s very nice to work on my current projects which we’ve set up our observability to modern standards when compared to any of the log vomiting services I’ve worked on in the past.

Obviously easier to start with everything set up nicely in a Greenfield project, but don’t let perfect be the enemy of good—iterative improvements on badly designed observability nearly always pays off.

Good tracing & monitoring means you should basically never need to look at logs.

Pipe them all into a dumb S3 bucket with less than a week retention and grep away for that one time out of 1000 when you didn’t put enough info on the trace or fire enough metrics. Remove redundant logs that are covered by traces and metrics to keep costs down (or at least drop them to debug log level and only store info & up if they’re helpful during local dev).

Hardware transcoding on SBCs is generally not fantastic, you’re gonna want to look for one that has VAAPI/VDPAU support or you’re gonna be looking at 100% CPU for half a day to transcode a film, which will make your other services effectively unavailable at the time.

I used to run my Plex server on a Pi4 with 4GB of ram and it basically crashed any time transcoding kicked in, I swapped to an intel NUC so I could get QuickSync for transcoding.

I’ll point out though, every SBC you’ve listed has usb, which is all you need for an external disk. If you’re worried about size, I’ve got a 5tb external drive that’s about 5cm², which is basically the footprint of any SBC you could use in this scenario

Okay fair play, if you’re doing this super short term it could make sense. Though I question what SBC you’re using that’s capable of transcoding video but not the ability to plug in an external drive.

$12/m for your 2TB of usage would make sense for maybe 5 months before it would be cheaper to buy an external disk—and of course that storage is gone once that time is up, Vs a hard disk which will probably last you a decade or so

I’m not sure about transparently, that’s more in the tdarr wheelhouse I’d say. You’d dump the files into a monitored folder and it will replace it with a version transcoded to your specification.

Transcoding video takes a fair bit of time and energy too FWIW, so you’re going to need enough local storage to handle both the full size and smaller one.

I have to question the idea though, cloud storage is always more expensive than local for anything remotely non-temporary, and transcoding a load of video all the time is going to increase your energy bills. If you have any kind of internet bandwidth restrictions that’s gonna factor in too.

I’d say it would be better to save up for a cheap external hard drive to store your video on. For a year’s subscription to a cloud storage service that would provide enough space for a media library, you could probably get twice the amount of storage forever.

Unless you’ve got raw uncompressed video, any kind of transparent compression like you describe is only going to cost you in energy bills for no benefit. Most video is already compressed with specialised video compression as part of the file format, you can’t keep compressing stuff and getting smaller files.

The alternative is a lossy compression, which you could automate with some scripts or a transcoding tool like tdarr. This would reduce the quality of the video in order to reduce the file size

Ventoy, as everyone else says, is your friend here.

Though I saw something similar in a video recently which I’m gonna call out for completeness, the IODD devices that let you change the image on the fly:

http://en.iodd.kr/wiki/index.php/Main_Page

Obviously not as cheap as a usb stick and ventoy, but a pretty cool alternative for those with the additional use cases

I love your specific example screenshot

“Hey is this Microsoft support? Yeah, err, so I’ve got this MANUALLY_INITIATED_CRASH error, can you help?”

“Have you tried… Not initiating…a crash…?”

Nice

So what’s the point of the EOD one?

I honestly see zero benefit in it unless it’s a 24h operation with a shift handover.

You get focus time?

Also, what the hell is the point in an EOD standup if you’re gonna have another one in zero working minutes?