I have no insider knowledge, but a look thought their changelog[1] is not exactly inspiring, especially if you consider that Heroku doesn't support things like HTTP2.[2] There are almost no real changes beyond updates to languages and the occasional (rare) minor change to the platform, which is not a sign of a healthy actively supported service.
At work, we spend a LOT of money with heroku every month. We really wanted to use Protobuffers (please let's not discuss that ;) ), which require HTTP 2 for the streaming stuff. So we talked with them and were like "heyy, we need this" and while they didn't explicitly say they weren't going to support it, the response basically boiled down to "yeah, i wouldn't be holding your breath" ;)
You can use protobuff without gRPC if that helps for your use case. Twitch has open sourced their HTTP1 protobuff client/server library: https://github.com/twitchtv/twirp
As a Heroku customer, I'd appreciate a migration guide from Heroku to Northflank in your docs. What are the equivalent services (nodejs dynos, worker, scheduler, postgres, papertrail) and what are the steps to move. Thanks
Had to look it up so here it is, from the TV show, The Expanse:
"Belter is a term used to refer to persons born in the Asteroid Belt or the Moons of the outer planets. An inner usually refers to someone who originated from the Inner planets of the Sol system, but usually meaning from Earth, Luna, or Mars."
No, it didn’t. PG made HN after funding and using Reddit.
Reddit had invite-only subreddits at that time and Joel Spolsky’s even had some traction. PG considered doing something similar but decided it wasn’t flexible enough for his needs and made HN.
But those "more expensive options" are completely different services, right? This is like "why should Lenovo market monitors when they're also selling their more expensive laptops?"
In my experience developing salesforce stuff in the past, forcing people into the wrong services while charging a fortune for the headache and lockin is their entire business model.
This is a major blow to some of us: the free version of Nightscout has been relying on Heroku and they had an insanely meticulously detailed guide to set the whole thing up. https://nightscout.github.io/nightscout/new_user/ I will need to ask them what's next :(
Of course, I could figure out hosting myself, I do have a server but I don't need any more tamagotchis :(
In theory it should. I don’t mind paying the $8/month fee so I do.
I mentioned the guide and my work on the NightScout discord. If there’s interest I could do from scratch instructions for those doing entirely new setups.
Note: I’m the person who did the migration and wrote the notes linked to above.
I thought they were in the other comment. The migration is pretty simple now that heroku only hosts the static part of the site and the data is in Atlas.
Oh hey I have a friend who’s child has diabetes and I make a little self-hosting personal computer. I’ll be adding nightscout for him asap - thanks for developing this. Usually self-hostable tools are fun toys - this is a nice change of pace.
Can confirm Periwinkle is a thing. It started a couple of years ago.
However, as far as I know, it's also done. The scope was massively reduced once they realised the complexity of what they are trying to do (which isn't sunsetting Heroku, but more moving it in the direction that Salesforce Functions has taken). In the end, Periwinkle essentially transpired to be the banner that you now see at the top of Heroku properties.
Saying that, I've been out a few months. I don't know what the current plans might be, and I do wonder if recent events might have SFDC evaluating what's next.
I just evaluated SF functions and it didn't seem too great, especially for the $$$. Especially since one could setup an AWS lambda with the newer function url + IAM user auth, where AWS sig4 is already supported by Salesforce's Named credential system.
https://render.com/ has been the closest thing to a replacement we've found (we've deployed HUNDREDS of apps to Heroku over the years, so this would be a massive loss for us as well)
Also shout-out to https://glitch.com/ which puts particular emphasis around community and learning. I use it anytime I'm preparing code for a class, workshop, demo, or open source documentation.
Good question, it is some epic internet lore. It was from Stewart Butterfield, founder of Slack.
He co-founded a game company Ludocorp in 2002 to create "Game Neverending", then pivoted its photo sharing functionality to become Flickr and sold it to Yahoo in 2005.
He then co-founded another game company Tiny Speck in 2009 to create "Glitch", and in 2013 spun out the chat tool the team built while making Glitch to become Slack, which IPO'd in 2019.
(Funny enough, one HN commenter sort of called it happening again[0])
I believe Stewart gave (or sold?) the Glitch.com domain to Fog Creek in 2018[1] and the game's site lives on at a new domain [2]. Some discussion from (Fog Creek's) Glitch's launch about the domain name change at [3]
They were on glitch.me for a couple years before acquiring the .com domain, but they're a relatively established organization as essentially an evolution of Fog Creek.
at one point i spoke to a CEO who had leased a 4-letter .com domain for 2 years for a million of dollar (i don't remember the specific amount) for their startup idea.
i imagine there are businesses which facilitate these types of transactions
The startup is still around. I don't know what the social protocol here is about things like this, but hint.com is still doing its thing... getting doctors and patients directly connected.
+1 for Vercel. Ran my first project on it years ago when they were young and I was stupid. It was so easy to get a front end hosted. I'm still stupid but a lot more experienced now, and still use it for spinning up side projects
Yeah, DigitalOcean is a great alternative. I've been using it for my personal projects for a while now. I think they have a $5/month machine that can do wonders for a hobby project.
They do, I even used that professionally at a previous position.
$5-10/mo box with Dokku installed gave us a heroku-ish platform that could be scaled up if needed. The main downside being that you have to manage it yourself, but it really was quite easy for the scale we operated at. Can't speak for larger systems. But if you need a pretty simple, quick-deploy, small-ish-load servicing then it's a breeze.
~10 years ago I had a web proxy on heroku to bypass internet filters. It was used by people in countries like China, as well as by kids (and teachers) in schools. (At one point my highschool blocked google!) Heroku's free tier was always good enough for the server, but I did pay for SSL for a while.
After it had been around for a couple of years, someone accessed a phishing site through my proxy and incorrectly reported to heroku that I was hosting a phishing site, so they froze my account to investigate. I was eventually able to show them that I wasn't actually hosting any phishing site, but then they said that they were planning on changing their TOS to disallow any form of proxies, so they were going to keep me frozen anyways.
If your server ip shows up in the wrong log you will get legal correspondence almost certainly for police to check whether you are the origin server. Caching would be another issue.
It's kind of sad that Heroku has been left to die a slow death. What it brought to the industry was so far ahead of it's time in setting up easy yet powerful deployments and infrastructure.
Sure, it was never cost effective at scale, but that wasn't the point. The product has been, and continues to be, very inspirational from a customer UX standpoint. I say this as someone that works in deployment and infrastructure tooling. No matter how many "next gen" deployment tools we build to take advantage of the most sophisticated deployment techniques, the pure simplicity and elegance of Heroku was something to behold.
While it's sad that Heroku seems to be on it's way out, it has undoubtedly moved the industry forward and there are a number of compelling options to replace it and this point. Excited to see where the industry moves from here.
I'm not a Salesforce or Heroku employee, I just orbit their space, do projects, listen to podcasts etc. so this is just speculation on my part.
Salesforce is rolling out "Salesforce Functions", like ASW Lambda but for internal Salesforce work. A function is already logged in to your Salesforce account and has access to your data. It's for things that would be hard to do in their proprietary Apex language in the platform.
It's possible they're taking the Heroku resources and staff and getting them ready for Functions.
I have no idea what the financials have ever looked like, but it's really sad and disappointing that a service that used to be so invaluable just gave up their lead and withered. There's definitely a lot of competition as Big Cloud offers more turnkey solutions (rise of serverless) and Small Cloud offers more power (Netlify, Vercel) so Heroku is squeezed on both sides. But they had the chance to fill all those niches first and they just didn't even try.
Yes I've heard this. Lack of product releases since 2019 and the significant reduction in headcount all point to that, still making $300m-500m ARR revenue so not something you should let die lightly...
I'm working on Northflank, a production ready Heroku alternative over at https://northflank.com
Builds, deploys, stateful workloads, advanced networking, crons, API and more.
I have a project that uses Heroku Postgres because they handle high-availability and backups, and because they have some fancy metrics that help find areas for potential optimization[1]. Does anyone have recommendations for an alternative with similar features? As far as I can tell RDS should handle the HA stuff, but it doesn't look like they have query-level metrics.
I'm open to looking at a separate dashboard of some kind for query metrics if necessary, but I pretty much need the reliability features to be abstracted away. This product needs to be reliable to be worth anything, and I handle development/deployments/infrastructure by myself so I don't have the required knowledge or time to build out a proper solution myself.
Crunchy Bridge is staffed with a lot of former Herokai and some Postgres core committers.
You can use PGAnalyze as a (fantastic) secondary service to give you statement/query visibility. I get the impression Crunchy is trying to build what Postgres data is (and what it could have been).
I do business with PGAnalyze but have only talked with Crunchy. I have heard good things about Aiven as well.
I mean I would understand if they wanted to do some rebranding, but even then I'd rename it to e.g. "Heroku by Salesforce" instead of sunsetting it entirely. It was a good brand, and any brand damage is of their own making (see other comments from insiders).
They didn't want to support Indian Cards for payment by adopting secure practices as per RBI guidelines. One of the big reason I couldn't continue on Heroku.
The unit economics of "free" work well if your product fits - if you have something that nets you a large operating margin, a good level of conversion to paying users after trying out the free tier (eg people have a reason to upgrade), that most people don't fully utilize their quota (eg you can oversell without killing the service), and, most crucially, is hard to abuse then free tiers can work brilliantly.
I have no idea if Heroku fits that model, but plenty of businesses run it very successfully.
I guess you bring a fair point, I didn't realize they were bought out till this thread. I'm guessing Salesforce is in it for the profits, which is probably fair.
...or, you know, on IaaP - to give the concept of having your own computer running under your own supervision on your own connection a fancy term. Infrastructure as a Property [1]. All you need is a computer - you're sure to have one of those laying around somewhere - and you're set. You'll be amazed at how much stuff you can host on that old laptop with the broken screen. It even comes with its own built-in UPS, imagine that! Add some external storage if needed and/or for redundancy and you're off to the races. Just make sure to set the thing to automatically install security updates and to make regular backups (rsnapshot configured for hourly snapshots with 3 months retention will go a long way here) and your stuff will be safe and secure - more safe and secure than when it is hosted at some big juicy target like Heroku.
Source: I've been doing this for more than 25 years. Never hot "hacked", never lost important data. I have seen countless drives and power supplies fail but always kept configuration and user-generated data safe (and that is all that matters, the rest can be easily re-installed from distribution media/the 'net).
[1] Maybe I should make a fancy content-less website with annoying scrolling habits for this to attract some VC capital
This comment is needlessly condescending, and you're already describing a ton of system administrator skills that you need to have, plus a good internet connection, plus hardware, which makes a lot of assumptions already.
> Never hot "hacked", never lost important data.
You got lucky. I'm not saying cloud providers are better, I'm saying you got lucky.
...or maybe the risk of "getting hacked" is not as big as it is made out to be given some simple precautions? I am not the only one who "got lucky" after all. Given an up to date distribution with only needed ports open to the 'net and a sensible password for those who use SSH password authentication you'll be safer than at most cloud providers. It is, after all, far more lucrative to try to gain access to the likes of Heroku than it is to JoeSchmoe.org.
Also, "needlessly condescending", give me a break. This site is called Hacker News so it is silly to call a call for exploration - the essence of the hacker spirit - "condescending".
I do not find it condescending. I find the perspective refreshing. Of course self hosting is not an option when building a product, you want to outsource what you do not expect to become an expert at and which is not part of your core business. Any form of operations and hosting quickly becomes such a thing.
Having said that I have also self hosted to 15 years. Arguably services that gave high utility, but never anything related to core business. I for one host everything on Digital Ocean. As a consultant I dont do enterprise cloud deployments very often, but when I do, I chose AWS and the client has the funding and pays for it.
Most people don't have internet connections fast enough to handle running their own website. Sure, you might handle a low-traffic page fine, but all it takes is hitting the second page of HN to bring it down.
Until about 2 years ago, I was on a 30 mbps connection. Gigabit wasn't even an available option.
https://fly.io has been amazing for me. Dunno how their docs used to look but IMHO today it is very detailed they are especially forward-thinking regarding documenting some nice edge use-cases for the platform.
Glad to hear that. Giving them another look is on my list, I could tell they cared about their docs but it was clear they were still filling in the gaps in some areas.
https://flightcontrol.dev gives you some sort of `your own Heroku on your own AWS acct` very interesting concept, also can be a replacement for Heroku.
they didn't buy heroku for its hosting, and my impression is that they've been somewhat frustrated by the fact that it keeps being so profitable. They bought heroku because of all the work heroku has done intergrating salesforce with other systems. I don't know much about that part of the business but yeah, they didn't really know what to do with the hosting side of heroku, and 100% would have killed it off right away if not for the fact that it keeps making good money.
[1]: https://devcenter.heroku.com/changelog
[2]: https://help.heroku.com/JAOCNZ25/does-heroku-have-plans-to-s...