The gist of this is that we are now permanently banning use of pinging services (third-party or built on Glitch) to keep apps on Glitch awake, as part of our efforts to stabilize and support the platform.
If you have further questions, we encourage you to reach out to us at [email protected]. We appreciate your patience and your support as we work towards improving the platform so you all can keep creating!
Most of the web apps here don’t have to be up 24/7 and glitch is more meant for that so I think this a good decision. However discord bots need to be up 24/7. To counter this you can also host a website on the same container for your discord bot if it’s popular and people visiting the website help keep the bot up.
We reserve the right to delete, suspend, or terminate your access to, or ability to use, any and all Services that we determine to be placing undue strain on our infrastructure. These changes were made in response to ping services on Glitch and our efforts to make the site more stable. You can read more about those efforts here.
Can’t we ping every 5 or more minutes to keep our website active?
This is a real shame, but if it has to be done, then so be it. However, please do consider moving custom/better looking loading pages up the road-map
Eddie
P.S. This will probably have a huge impact on the Glitch Pro plan. So I guess Glitch will benefit from this, but at the same time loose a huge part of the community.
Well it wouldn’t be good to have a bunch of processes that aren’t doing anything. Unless you have a popular app(then you wouldn’t need a ping service to keep it up the users will help keep it up by visiting it), you don’t need to keep the site up when no one is using it
I hate it when something like make8bitart.com doesn’t load because other stuff that isn’t really being used(small websites, discord bots that get used only every once in a while) stopping it from loading.
I mean @jenn, you do understand people will still try to get around this. We have seen a huge amount of posts from people who have managed to get around the block. I think this will be fine for Web projects, if you added nicer loading pages, but Discord bots, which were what made Glitch so much popular will be gone. Hey, ho! Guess my bots are dead!
I would be fine with this change if the time that apps stayed idle without sleeping for was a bit longer. But this change is definitely turning me away from Glitch, which is sad. The pro plan just doesn’t compete with cheap VPS solutions, despite the ease of deployment.
I can’t see myself continuing to use Glitch for much longer, it’s been rather unreliable in lots of ways. I can’t complain, after all Glitch is a free host - and was one of the best.
Same. Why does it need to be 5 minutes? I understand it’s to keep Glitch fast for everyone, but can it be increased now that pinging services are blocked?
Exactly, I’d understand why it would be necessary for intensive applications. But for idle web apps, having to wait for it to wake up after 5 minutes being idle is frustrating.
It should be determined by CPU processing time, not a set time.
Why can’t it be 5 minutes. 5 minutes is enough for a user to explore your site. I don’t see why a user would take more than 5 minutes between nagivating your sites. If you have long content you can just split it into pages
That indeed is a good thing, at least then people will consider buying the pro plan, and then get income. But at this point I can already see a popularity graph falling, fast.
Nah Glitch is too ease of use, and easy to deploy. I’m sure Glitch will only suffer a minor loss because of this. We will lose most of the discord community though.
Buy the pro plan then. Glitch is designed for websites not discord bots. If you changed the timeout to something like 1 hour, then there would be a lot of websites up not doing anything and it would make glitch’s resource bill a lot higher.
Glitch will still be a great platform for developing Discord bots and seeing near-instant results of changes. I never saw Glitch as a great place to host Discord bots anyway, I’ve always stuck to using it for APIs and hobby projects.
A lot of users use Glitch for Discord bots, I would be interested to see how that impacts Glitch. I hope that the webdev community will discover Glitch.
Hey there - I totally understand the frustration and encourage you to use whichever platform best suits your needs – and I totally understand that Glitch may not be the one right now! Hopefully it will be again in the future for you. All I ask is that while you’re in this forum you be respectful and kind to (and about) others.
I haven’t really had my ear to the ground on infrastructure discussions at the cloud resource level, honestly. But having moved from one service to another at a previous job, it’s not a simple or cheap task!
The company I’m working for moved from AWS to self-hosted (we bought a bunch of servers and stuff and set it up in Germany, that was fun). If it wasn’t for the flight tickets, hotel rooms and transport we’d nearly not pay anything. We have co-location with Contabo. Moving all of our services was just a bit time consuming, but not really expensive.
I’m sorry, but I don’t really see how moving from AWS to another hosting provider will be all that expensive? Can you clarify?
Well I am moving my bot on my phone to host it 24/7 there. At least I won’t have any limits except my phone’s limits. I used Glitch since it is easy and fast to deploy, but since bot will go offline after 5 minutes I can’t host on glitch anymore. I can think of developing program to do remote change deployment easy and fast. I will host only websites but besides that for example like bots, I will need to self-host.
I understand the decision and I’m completely fine with it. Glitch isn’t a platform to host stuff 24/7. It’s a community where you can share your projects, work on them or test them. Personally I’m okay with waiting for projects to start up (if they no longer get stuck), because hey - it’s free!
In my opinion, this is a good and bad idea… a lot of people use glitch for discord bots which disabling pinging services would kinda make a lot of people who host discord bots will stop using glitch because there are probably a lot of kids using glitch for discord bot hosting and they need it 24/7. It wouldn’t really increase more people to buy boosted apps because there’s a lot more cheaper discord bot hosting providers out there. beat me i actually use a host to host one of my discord bots
Hi! I am very happy to hear that pinging services are banned. But reading the original post and some parts of the discussion, I am confused about why would someone even need pinging services if the projects are awake? Am I getting something wrong?
I’ve been using Glitch for a couple of years now, pretty much since it got the Glitch name (after Gomix + Hyperdev) for little experiments and mostly to teach other people to code. I absolutely love the product and the team.
Recently it was frustrating to recommend Glitch to people to make their first HTML site only for it to be flaky/down, but it turns out this was all because a chunk of users were abusing a free system to make always-on bots.
For anyone who didn’t already “get it” from the great replies by Cloud, r0ss, and others, this is something called the tragedy of the commons.
If you have paying users but others are getting the primary paid benefit (always on) by using a hack, then why pay? It becomes impossible to create a paying user base that will sustain the service for everyone.
As it happens, I probably won’t pay for Glitch, but you have to recognise that the people who pay need to have a reason to do so, and Glitch needs to earn more than free-user goodwill - it needs revenue.
So I’m personally glad about the ban and I hope that this opens a new prosperous era of committed paid users and frugal, satisfied free-users like myself.
Cloud computing is very expensive. Lets say that Glitch has 2000 projects running and online at any one time, (which is a lot, but its possible). Lets also say that 50 containers can fit into one VM. According to the AWS Pricing Calculator, this would set you back around 30,000 dollars per month. That also isn’t including the s3 Glitch probably uses for cdn.glitch.com.
Thats also assuming Glitch uses that few vms. According to @cori:
I absolutely understand the reasoning behind this and think it’s the right move, but I can’t help but feel like this is something that could have been communicated via email.
I went out of my way to confirm that I WAS allowed to keep my app alive according to the documented limits before I built anything. Now that page 404’s and today when I needed to use one of my apps they were failed without explanation.
Googling glitch + “limits” or “restrictions” or “keep alive” don’t bring up this announcement so the only way I figured out the problem was complaining to friends.
If I’d gotten an email saying my apps were going to fail I’d have just upgraded to a pro plan and been happy to do it. Now I’ll almost certainly upgrade, but grudgingly so, and with a lot of concern that I can’t treat Glitch as dependable for anything other than toys.
Basically what happened here was, “Hey we have a new service that provides free french fries, it’s open for all in the mall’s parking lot” and it was a great service, plenty of fries, all day, all night, everyone loves the fries. And then eventually the service went “Why are there so many seagulls here???” and they’re putting up nets on top of the parking lot in order to stop the seagulls.
Could there be a sys where if user1 clicks some button allowing others to use their servers then user2 pings and based on servers available you could return successful ping or not successful ping? All my apps don’t need to stay awake 24/7