5 ms·

Does anyone have a mirror? The link is down. Nothing against the OP, but I don't know how in 2021, websites can't handle a few dozen visitors at once. Even a $

Does anyone have a mirror? The link is down. Nothing against the OP, but I don't know how in 2021, websites can't handle a few dozen visitors at once. Even a $5 digital ocean machine can handle thousands simultaneously without going down.

keane 6y ago
http://webcache.googleusercontent.com/search?q=cache:https://mybyways.com/blog/what-i-learnt-shopping-for-decent-computer-speakers http://webcache.googleusercontent.com/search?q=cache:https:/...
naavis 6y ago
Getting to the front page of HN brings a lot more traffic than "a few dozen visitors at once". https://nicklafferty.com/blog/what-happens-when-you-re-on-the-front-page-of-hacker-news/ https://nicklafferty.com/blog/what-happens-when-you-re-on-th... https://thehftguy.com/2017/09/26/hitting-hacker-news-front-page-how-much-traffic-do-you-get/ https://thehftguy.com/2017/09/26/hitting-hacker-news-front-p...
- LukeShu 6y ago
  Per the screenshot in latter: 1571 page views per hour. That's 26 page views a minute, which is "a few dozen".
  - naavis 6y ago
    The screenshot is from the point the email was sent, not the point where the site went down.
- amelius 6y ago
  18.000 visitors per day is 1 visitor every 5 seconds.
  - naavis 6y ago
    When you post stuff on HN, the visitors don't usually come in an orderly queue at the same rate for 24 hours. There is most likely a spike where most of the visitors open the site within a couple of hours, and then click around the other parts of the site. It can get pretty taxing pretty fast if you are serving for example images, and each page view triggers database queries.
- HenryBemis 6y ago
  We sometimes call his "the HN kiss of death", and we do this a lot to smaller visibility websites that never anticipate a few hundred/thousand hits within a couple of hours. I usually go through the comments, then drop these URIs on my calendar and check them a couple of days later. By then the discussions have matured and the site is up and running (perhaps even the posts have been correct/enhanced with the HN comments).
geocrasher 6y ago
Well, the OP isn't using such a thing. A quick lookup says it's hosted at internet-webhosting.com which appears to be a Malaysian web hosting outfit that sells the usual shared/vps/dedi hosting along with "wordpress hosting" which looks a lot like cPanel + Litespeed. Ho Hum. https://web.archive.org/web/20201026180211/https://mybyways.com/blog/ https://web.archive.org/web/20201026180211/https://mybyways.... shows that it's built using Grav, which is fine I guess. I don't know how well that holds up to high traffic, but if he's not using any specific caching then I think just about anything could fall over in the face of the HN hug of death. And until one has been HN'd (or slashdotted) then they really don't think about such things most of the time. I know I didn't. Then, my WP site got slashdotted. It stayed online with some quick work, but that was almost a decade ago. I suspect that the OP is doing a similar scramble now.
- foepys 6y ago
  I'm getting a "connection was reset" error. The hoster is most likely blocking all requests to the host name, probably due to an automatic DDoS detection (Wordpress is quite memory hungry) or because the included traffic has been exceeded.
kbenson 6y ago
Because some things don't even show up as a problem until you've encountered them before. I'm not sure about now, but Wordpress used to be nigh-unusable for a blog if you got any sort of notable traffic unless you you installed and configured a caching plugin. People that went through that generally did it from then on for any future blog they configured, but that doesn't mean some new blog from some new blogger would have it, or that someone wouldn't forget about it when setting up a new one. The problem is that it's often not a slow degradation, it's an exponential one. Things will be going fine, and then things will look a little slower, and then things will completely fall apart. This is because the few parts that actual lock and cause contention stack, and problems beget more problems. See also virtual memory in servers. Often, that point at which your server needs virtual memory is the point at which it slows everything down more that more and more requests stack up (because virtual memory is slow), needing more memory, and you get a spiral of death in many cases.
m463 6y ago
https://web.archive.org/web/20210321221948/https://mybyways.com/blog/what-i-learnt-shopping-for-decent-computer-speakers https://web.archive.org/web/20210321221948/https://mybyways....

keane 6y ago

http://webcache.googleusercontent.com/search?q=cache:https://mybyways.com/blog/what-i-learnt-shopping-for-decent-computer-speakers http://webcache.googleusercontent.com/search?q=cache:https:/...

naavis 6y ago

Getting to the front page of HN brings a lot more traffic than "a few dozen visitors at once". https://nicklafferty.com/blog/what-happens-when-you-re-on-the-front-page-of-hacker-news/ https://nicklafferty.com/blog/what-happens-when-you-re-on-th... https://thehftguy.com/2017/09/26/hitting-hacker-news-front-page-how-much-traffic-do-you-get/ https://thehftguy.com/2017/09/26/hitting-hacker-news-front-p...

LukeShu 6y ago
Per the screenshot in latter: 1571 page views per hour. That's 26 page views a minute, which is "a few dozen".
- naavis 6y ago
  The screenshot is from the point the email was sent, not the point where the site went down.
amelius 6y ago
18.000 visitors per day is 1 visitor every 5 seconds.
- naavis 6y ago
  When you post stuff on HN, the visitors don't usually come in an orderly queue at the same rate for 24 hours. There is most likely a spike where most of the visitors open the site within a couple of hours, and then click around the other parts of the site. It can get pretty taxing pretty fast if you are serving for example images, and each page view triggers database queries.
HenryBemis 6y ago
We sometimes call his "the HN kiss of death", and we do this a lot to smaller visibility websites that never anticipate a few hundred/thousand hits within a couple of hours. I usually go through the comments, then drop these URIs on my calendar and check them a couple of days later. By then the discussions have matured and the site is up and running (perhaps even the posts have been correct/enhanced with the HN comments).

geocrasher 6y ago

Well, the OP isn't using such a thing. A quick lookup says it's hosted at internet-webhosting.com which appears to be a Malaysian web hosting outfit that sells the usual shared/vps/dedi hosting along with "wordpress hosting" which looks a lot like cPanel + Litespeed. Ho Hum. https://web.archive.org/web/20201026180211/https://mybyways.com/blog/ https://web.archive.org/web/20201026180211/https://mybyways.... shows that it's built using Grav, which is fine I guess. I don't know how well that holds up to high traffic, but if he's not using any specific caching then I think just about anything could fall over in the face of the HN hug of death. And until one has been HN'd (or slashdotted) then they really don't think about such things most of the time. I know I didn't. Then, my WP site got slashdotted. It stayed online with some quick work, but that was almost a decade ago. I suspect that the OP is doing a similar scramble now.

foepys 6y ago
I'm getting a "connection was reset" error. The hoster is most likely blocking all requests to the host name, probably due to an automatic DDoS detection (Wordpress is quite memory hungry) or because the included traffic has been exceeded.

kbenson 6y ago

Because some things don't even show up as a problem until you've encountered them before. I'm not sure about now, but Wordpress used to be nigh-unusable for a blog if you got any sort of notable traffic unless you you installed and configured a caching plugin. People that went through that generally did it from then on for any future blog they configured, but that doesn't mean some new blog from some new blogger would have it, or that someone wouldn't forget about it when setting up a new one. The problem is that it's often not a slow degradation, it's an exponential one. Things will be going fine, and then things will look a little slower, and then things will completely fall apart. This is because the few parts that actual lock and cause contention stack, and problems beget more problems. See also virtual memory in servers. Often, that point at which your server needs virtual memory is the point at which it slows everything down more that more and more requests stack up (because virtual memory is slow), needing more memory, and you get a spiral of death in many cases.

m463 6y ago

https://web.archive.org/web/20210321221948/https://mybyways.com/blog/what-i-learnt-shopping-for-decent-computer-speakers https://web.archive.org/web/20210321221948/https://mybyways....