EDIT: there are actually currently existing methods that users can use to send api requests which will send back server data. but these are rather obscure and do not communicate with the average user very well. this does prove, however, that there are already resources that could make an implementation of a much more formatted website easier.

it’s no secret that scratch has had server issues over the past few years, and unfortunately there’s nothing we can do but wait it out. however, communicating with your userbase that your website has temporarily went offline, primarily due to site traffic or updates, adds clarification to the issue and won’t have people nearly as confused.

right now, when the scratch website goes offline, the only way we can confirm the site either went down due to network issues or for manual maintenance is through scratch’s social media page.



this is a fine approach, but it could be way better. sometimes not even the entire site goes down, more often than not it’s actually separate parts of the website! plus, there’s no guarantee there will be a post from scratch’s social media account that something happened on the site (mainly because it’s something that happens pretty fast).

this is where the server status website comes in. when the site goes down, instead of redirecting just to scratch’s social media page, it will also redirect you to the server status site, which may contain information on various stuff such as…

• front page project rows
  
• messages inbox
  
• cloud variable server
  
• various 2.0 pages
  
• various 3.0 pages
  
• comment sections (on rare occasions they will manually go offline)
  
• discussion forums
  
• etc.
  

whether or not the user is a developer, the status page will benefit pretty much anyone and everyone. as it lets them know why something on the website has gone wrong. you could be, let’s say, a teacher who wonders why they are unable to setup student accounts for their class. when the teacher checks the status website, they realize that the account creation system may be having issues. this is just one niche example of many others you yourself can think of.

to you scratch geeks out there: please correct mistakes i do make in my writing.

support, the @Scratch bluebird account hasn't been active in a while, and if the maintenance is not announced people can still know what is wrong with the website.

Support! Helps keep scratch isolated from bad websites by not referencing to them.

Hypothetically; If the server ever goes down, would the stats site not go down with it?

starlightsparker wrote:

Hypothetically; If the server ever goes down, would the stats site not go down with it?

bruh just use different servers for the stats website

I support as Scratch is for 8+ and many 8 year olds wouldn't have access to X*.
*It doesn't have Skibidi Toilet on it.

8to16 wrote:

bruh just use different servers for the stats website

One could assume that could potentially be costly.
And is there a possibility that MIT’s side ever causes issues?

TheInterneter wrote:

*It doesn't have Skibidi Toilet on it.

wdym X does have skibidi toilet
(every social media at least has at least some post with skibidi toilet at this point)

8to16 wrote:

support, the @Scratch bluebird account hasn't been active in a while, and if the maintenance is not announced people can still know what is wrong with the website.

It is active, but not many people can access it when needed

starlightsparker wrote:

8to16 wrote:

bruh just use different servers for the stats website

One could assume that could potentially be costly.
And is there a possibility that MIT’s side ever causes issues?

Not really.
I’ve ran servers before which took barely anything more than running a system normally

GlitchedThrough wrote:

starlightsparker wrote:

8to16 wrote:

bruh just use different servers for the stats website

One could assume that could potentially be costly.
And is there a possibility that MIT’s side ever causes issues?

Not really.
I’ve ran servers before which took barely anything more than running a system normally

I mean, it's basically a normal computer, just designed to be connected to via other computers and without a display (usually).

SpyCoderX wrote:

GlitchedThrough wrote:

starlightsparker wrote:

8to16 wrote:

bruh just use different servers for the stats website

One could assume that could potentially be costly.
And is there a possibility that MIT’s side ever causes issues?

Not really.
I’ve ran servers before which took barely anything more than running a system normally

I mean, it's basically a normal computer, just designed to be connected to via other computers and without a display (usually).

I mean, anything with Internet connectivity is designed to be connected via other computers as well

starlightsparker wrote:

Hypothetically; If the server ever goes down, would the stats site not go down with it?

That is an existing problem with scratch.mit.edu/health

starlightsparker wrote:

8to16 wrote:

bruh just use different servers for the stats website

One could assume that could potentially be costly.
And is there a possibility that MIT’s side ever causes issues?

Use one of the many other domains that the ST owns (so they don't have to use MIT's servers, or anything else). They already own all of

- scratch.love
- scratch.computer
- scratch.chat
- scratch.rest
- scratch.team
- scratch.pizza

some of which redirect to scratch.org, while others used to redirect to the homepage and now throw a 403 with a message saying something along the lines of “Hissssssss! Markley VPN Only”.

Or, they could use some trickery related to why @codubee's profile picture is actually a random pick out of 3 different ones and have Fastly actually be the ones hosting the server status page, at the cost of updating less frequently and probably being less dynamic.

edit: formating

BigNate469 wrote:

some of which redirect to scratch.org, while others used to redirect to the homepage and now throw a 403 with a message saying something along the lines of “Hissssssss! Markley VPN Only”.

I wonder why they hiss at you for not using Markleythis is my derail shield. i will use this when anyone derails the topic. do not remove this.

bump

To anyone asking if this would be costly, it definitely wouldn't; health checks are simple and cheap to do. This could extend further to identify when the servers are under significant stress so Scratch developers can know what is causing bottlenecks

(unless they are already doing this internally which would be great)

ajskateboarder wrote:

(unless they are already doing this internally which would be great)

i think their server provider sends active server info to the scratch team on a regular basis. we can prove this by the tweets scratch makes during outages

LaughabIe wrote:

ajskateboarder wrote:

(unless they are already doing this internally which would be great)

i think their server provider sends active server info to the scratch team on a regular basis. we can prove this by the tweets scratch makes during outages

Well then! That would be nice to also show as regular messages on the status website instead of as tweets

TheInterneter wrote:

*It doesn't have Skibidi Toilet on it.

X most certainly does have Skibidi Toilet on it, because I think the creator has an X account

8to16 wrote:

BigNate469 wrote:

some of which redirect to scratch.org, while others used to redirect to the homepage and now throw a 403 with a message saying something along the lines of “Hissssssss! Markley VPN Only”.

I wonder why they hiss at you for not using Markleythis is my derail shield. i will use this when anyone derails the topic. do not remove this.

Markley is a VPN that all ST members are required to use. scratch.pizza and scratch.love are used as staging domains if I recall correctly- and thus are designed to only be usable by ST members. Since Markley seems not to be public (the ST made it, I think), if you have Markley detection you effectively have a ST detector.

I don't know why they didn't just make it redirect to the homepage for anyone without Markley, though.