Show HN: Server Uptime
30 points
by dIsK
3 days ago
| 11 comments
| server-uptime.pw
| HN
progbits
3 days ago
[-]
Are long uptimes still a thing? Both at home and in production at work I instead strive for regular updates and restarts.
reply
stackskipton
3 days ago
[-]
I would imagine not anymore. I generally see them as a sign of serious tech debt at this point.
reply
iforgotpassword
3 days ago
[-]
Reminds me of the long-dead uptime-project.net

On their leaderboard there was an NT4.0 machine with several years of uptime. People accused the owner of having manipulated the reporting tool, but the owner said it was just being used as a secondary DNS server internally at a company, and scheduled to be replaced as soon as the hardware fails. I don't know whether that happened before the site shut down unfortunately.

reply
donatj
3 days ago
[-]
I 100% believe it. I had a Windows 2000 smb/file server run for over 3 years uninterrupted until there was a power outage.
reply
alexwasserman
3 days ago
[-]
A while ago at a company selling monitoring software I deployed an updated agent out to clients and immediately a large number of their Solaris (SPARC) servers went orange or red. Pretty quickly worried about what I'd done.

But, it turned out there was a bug in our earlier agent version that wrapped the uptime at 365 days and all their servers were well over that. They'd had warnings set at a year and critical at 18 months/2 years, some high value.

I asked if they were worried, and apologized for the bug. They said it wasn't a problem and just increase the thresholds so it was all green again and they'd get around to rebooting at some point.

reply
Gys
3 days ago
[-]
Why would I send the uptime to a remote server?
reply
whatevaa
2 days ago
[-]
To tell that there is high likehood that your server doesn't have some kernel patches applied
reply
KetoManx64
3 days ago
[-]
For the feeling of superiority of course, why else?
reply
rossant
3 days ago
[-]
This one is great: https://healthchecks.io/
reply
theideaofcoffee
3 days ago
[-]
Second healthchecksio. Used them recently for a quick project that “just” needed status monitoring for cron tasks. Super simple to drop a curl equivalent at the start and end and important moments in the middle. Do one thing well, yadda yadda. This one seems pretty similar in scope to hcio where you can also send simple (simple!) payloads like this project as well.
reply
rozenmd
3 days ago
[-]
As a fellow uptime monitoring vendor I couldn't figure out what this does - an explainer paragraph would do wonders
reply
bubblesnort
3 days ago
[-]
The explanation is right there on the page.
reply
johannes1234321
3 days ago
[-]
The page explains how to send some data collected by some program to some host, but not what exactly that data is (uptime duration only or also load information or identifying information?) for what purpose and what that server does with it.
reply
xist
3 days ago
[-]
I still think of the freebsd server with just under 13 years of uptime when I killed it. No one can explain the rush of happiness that comes with issuing a shutdown command.

I tried powering it back online afterwards. It did not power back on.

reply
freitzzz
3 days ago
[-]
Funny. Server Uptime but the host isn’t reachable
reply
johnklos
3 days ago
[-]
Ah - it's a project that's only for computers that use apt, and probably only amd64. How uninteresting.
reply
codetrotter
3 days ago
[-]
https://www.freshports.org/sysutils/uptimed

The uptimed package they told you to install also exists on for example FreeBSD. And from the table it looks to be available on platforms like aarch64 (64-bit ARM), and powerpc too. (Unsurprisingly, given the very straightforward-to-implement thing it does.)

So to answer your statement: No, and no.

And all the program does is keep track of the highest uptime your system has ever had. And the website takes these records from users, and presents the submitted data online.

I dont think the website is even made by the same people, probably.

https://github.com/rpodgorny/uptimed for the uptime max stuff that runs on your machine, and all this website asks you to do is to submit your high scores from that other tool.

reply
johnklos
3 days ago
[-]
Thank you. This makes much more sense.

Whoever made that page should've had some better information.

reply
putna
3 days ago
[-]
i wonder if ‘finger’ and ‘.plan’ files can be used for uptime monitoring in similar fashion
reply
rc_kas
3 days ago
[-]
you should probably open source this one.
reply
codetrotter
3 days ago
[-]
The tool that they tell you to install is https://github.com/rpodgorny/uptimed which is open source

Are you asking the owner of this website that takes data submitted by users to open source the website? Or are you talking about the above-mentioned program you have to run, which is already open source?

reply