No deployment provides persistent storage. To achieve this, you need to connect a database of some kind. Replit staff is aware of this and is looking into persistent storage options for deployments.
right, that’s what i heard day one. how and when will we be informed of the mighty oz’s decision on the matter? as the clock ticks, there is a lot of work to be done with both the repl.it domain going away and needing to figure how that part will be working on 01.02.2024 as well as refactoring existing code to work with a db and working out if the replit db will work for the project or if an external connection will be required (and of course then incurring additional egress charges).
tbh, i’m already getting the majority of these projects up and going elsewhere at this point, but am for whatever reason, remotely hopeful there is a way i can keep something if not everything here as i remain fond of the community.
Thanks for the good news @soren but to help @craigz and probably many people including me, do you mean that the object storage solution will act exactly as a Linux filesystem like today or like a remote object storage lik GCP cloud ?
When you announced stopping always on I started deploying and I’m now happy with deploys but accessing gcp cloud storage is not as easy as having a real file system and it needs programs to be updated. And it is was far from normal FS that i prefer now to switch to Postgres
I hope it will be something really good for users that loved simplicity of Repl
I share the sentiment of surprise and disappointment with the recent announcement from Replit. It’s understandable that changes and improvements are part of the growth process, but it’s crucial to consider the impact on the user base, especially the younger demographic.
I resonate with the concern about introducing premium subscriptions and the potential challenges it may pose for the teen user base. Many of us, being part of this group, may find it challenging to access premium features without relying on our parents’ cards. It would be great if Replit could explore options that allow for free deployment with certain restrictions, making it more inclusive for all users.
Regarding the suggestion by @Classfied3D about differentiating between 24/7 repls and request-based projects, it’s essential to consider that Replit serves a diverse user base with various project types. Treating all projects equally ensures fairness and a level playing field. If there are limitations, they should be applied universally and not favor one project type over another.
Let’s brainstorm and come up with constructive ideas to address these concerns and maybe together we can request them .
I think this is a terrible idea. Correct me if I’m wrong, but half of the Replit users are here for the free hosting. I don’t care if this is better quality. The way I see it, this is just another corporation trying to make money. Do you care about the Replit users? You say you do, however this has to be hurting your popularity. I’ve seen tons of comments saying people are moving to services like Glitch.
And honestly, if you can’t provide even slow and bad connections to users for free, why should I pay for a plan or deployment? And yes, I know the Static deployments are free, but if you go over the Egress limit, it charges you money, when it really should be canceling the deployment until you CONFIRM that you want it to charge you money.
I’m devastated to see the amazing service that Replit was being deprecated. Sure, I may get a Hacker plan or pay for a deployment, but I’m 100% sure that you will lose a lot of activity here.
Goodbye to the good ol’ days, Replit. It was fun while it lasted.
I’ve just watched a youtube video, and it strikes on an important point. Sometimes, change has to happen. I’ve been kinda all over the place in terms of opinions, but I’m only going to use my emotions in moderation.
When a company starts up, they want a userbase. So, such companies would offer unbeatable plans (such as free hosting…) while losing a ton of money. This isn’t exclusive to Replit, as hop.sh as recently transitioned out of this phase (just like heroku) and uber has been constantly losing money until some time recently.
Then, once their userbase is big enough, they can’t sustain as a free service, so that’s why they need to add money. I acknowledge that the Replit team’s execution wasn’t great, but sometimes a company just has to be a company.
— DISCLAIMER —
I still oppose the change, and this post does not refute or challenge that fact in any way.
Indeed, the landscape of Replit is transforming, and the recent shift towards a more commercial model has stirred reflections among the community. The origins of Replit, rooted in fostering creativity and providing a haven for learners, drew users in with the allure of free hosting and an inclusive environment. It’s this very ethos that resonated with many, including myself.
Now, as the company traverses a path toward sustainability, the challenge is to preserve the spirit that made Replit unique in the first place. The introduction of paid hosting raises questions, not against the need for change, but about the preservation of the fundamental ideals that users have come to appreciate.
If that happens , there will rise one more Replit … maybe GLITCH is the new one…?!
This blog post says they will two more modes with ‘flexible pricing’ (on demand and cron jobs). On demand will have a free tier (I imagine it will be like wake-up repls without pingers). Cron jobs sounds like a timed thing (e.g. run a task every 5 minutes). But for some reason I’ve seen nothing else about this.
Essentially, but they will probably limit uptime to a small amount so it can’t be abused. I imagine we will use on demand for small, not very powerful free hosting such as APIs, static for static hosting, autoscale for better server-usage hosting and uptime, reserved vms for 24/7 repls and cron for tasks which need run ever so often
Pro does not still exist, it’s down to just core. Also, it being 10$ a month is only a promotion. It goes back up to 20$ a month at some point. (I think. I know the yearly will go back up to 220$ from 120$)