I knew when I got hit with the per seat pricing change a few months ago that there were preparing a raise and needed to get those numbers up. Good for them I guess.
Funny enough I had big VC call me for due diligence (because I posted a lot on their forums and they spotted my name) and I happily told them how pissed I was with that change and how I went from vocal evangelist for the platform to looking for an alternative.
That change really sat poorly with me as well. My brother and I have a small pet project that makes zero dollars but we would have had to pay $50/month just to allow him access to the project on Render.
Still not as _terrible_ as Vercel which straight up doesn't deploy if the user isn't a paid org member lol. Now that is shameful as shit!
But this pricing change from Render was bad yeah. :(
I mean I would have accepted it if they had network separation for various environments managed by the same team, but they DON'T (anyone telling you its fine to deploy staging in the same network as prod should be fired IMO). So you have to use different teams and pay X times the seat pricing which is ridiculous.
Hi folks, I'm the founder and CEO at Render. With this raise, the company is doubling down on making the cloud delightful for growing software teams who don't want to think about infrastructure.
We're looking for great people to join our team! See render.com/careers for all open roles.
As far as suggestions go, I moved from Heroku to Render but I'm again looking for an alternative. There was something that Heroku had that I'm strongly missing in Render: shared credits, for projects that have low usage.
I have _many_ tiny-usage projects, I'll say "10" but really many more. The thing is, even if you put their usage (resources, network, disk, etc) together it's still a "small" project equivalent, something that could run (capacity-wise) on a single VPS easily for $20-30/month. But the way Render is structured, even if I get 10 requests/month on each of them, I need to pay that per project. Which adds up quickly when you have many small projects as a hobby.
So I'm wondering if something like what Heroku had where there's a shared pool of 2000 credits that I can use among my projects as they get used by people, and when one of them becomes more popular then I can "upgrade" it to have a proper full-instance for each of its parts while keeping the rest of them in the pool.
If I wasn't so bad at devops, here the equivalent is I'd rent a mid-size VPS and have one folder/project, and I would be able to have them all hosted in that server with a lot of capacity to spare for $20-30/total. The way I have it now, if I wanted to host the same thing with Render, it'd be $250+/month only for my active projects. Heck, I think even my Raspberry Pi would be able to run all of these projects of mine properly (again, capacity-wise, I def love Render for its easiness of use).
I had the exact same problem with Render, but I wouldn't expect them to handle this situation because I'm guessing it's kind of an edge case? I ended up using a DigitalOcean Droplet + CapRover to do this, and then wrote a giant article detailing all my steps in case that's useful: https://alldjango.com/articles/serve-multiple-django-sites-f....
Yep, same situation. Looked into render. I liked what I saw (and spun up a trial). But it was going to be expensive for a bunch of low usage projects.
I ultimately went with a large-ish droplet from DO with dokku installed. It’s been great, for the many small sites use case. It doesn’t scale, but that’s fine. When projects get traction the intention is to move them.
Which one has, and is now on AWS. But if we’d been on render I wouldn’t have moved it. I’d just spin up its own db and scale up the processes.
Second, have you guys had a chance to revise your hiring pipeline? When I spoke to your recruitment team a bit over a year ago they were asking for 8 straight hours on a zoom call as part of the software engineering interview. I had to drop out of the process because fitting that in was exceptionally difficult. (They did offer to split it into 5 and 3 but I still couldn't fit that in). They remarked they were looking to update that process a bit.
The above aside - I remember thinking everyone I spoke to there was pleasant and enjoyed working there. I was sad to have to drop out.
Thank you. I agree that ask was a bit much for a lot of candidates; we updated our engineering interviews to be shorter and more manageable a few months ago.
I was very impressed by render when I was investigating hosting platforms, but if I'm being honest I got scared away by the postgres pricing(free for only 90 days: https://render.com/docs/free#free-postgresql-databases).
Question: After some more research, I noticed that even among competitors there aren't really any "always free" options(e.g. even AWS directly only offers 1 year free for RDS). What is it that makes free-tier postgres so uncompetitive in comparison to hosted compute? Is it because dedicated disk is more expensive than shared cores?
I'd like to give it a shot but I often don't get past HR because I'm in (theoretical) CS and lack "real world experience". Which is lame because the stuff I've done from my trench is amazing (source: trust me, lol).
Anyway, I'm really good on the backend, creating high performance REST APIs, documenting them, etc... and know a great deal of postgres.
I also have a good intuition for what's going to stick or not in the field. I was doing AWS ~15 years ago, back when you had to use an ugly SOAP interface to get things done, but I saw the potential and created the first automated cluster of bioinformatic tools in there. Right now, I'm betting on Rust, WASM, V8 isolates, and everything that has to do w/ lightweight function execution.
If that sounds good for render's mission, I'll be glad to take part in it.
Render's second engineer (who's still here and doing really well) had no prior professional experience in software and was entirely self-taught. I encourage you to apply!
Do you have thoughts on the Heroku->Render dynamic? Is there a name for it? I've seen it enough now to recognize a solid pattern for a successful business:
1. Company starts to do X, makes it a wonderful experience
2. Users love it, it's successful
3. Company stop treating users well
4. New company steps into make X a wonderful experience again, with the advantage that the market is pre-validated.
How much of an advantage is it to know that users want what you're creating? Do you have any tips, like ways to reassure people you won't follow the same dynamic as the original company?
That's an interesting thought. Any company that raises money from VCs will ultimately either fail or end up publicly traded (via IPO or acquisition), and at that point it's just a matter of time before they start squeezing their users.
I wonder what the ideal time is in a company's arc to start their eventual replacement. Render just raised a series B which probably means they're still many years away from step 3, so it's probably too early. But maybe when they're raising a series C or D, it's time to start thinking about making their replacement.
I feel like Stripe is entering that territory right now. Not that they're worse than alternatives, but they no longer have that "wonderful experience" magic because they've started to turn on the maximize shareholder value engine.
A subtle distinction is that it's not Heroku treating users poorly — it's Salesforce, and it's not the same people who were at the company even just ~2 years ago.
With Render, I'd rather show than tell. As an interesting data point, we've now existed as an independent entity longer than Heroku did, and we're truly just getting started.
(Render CEO) We're going to be hybrid and multi-cloud for the foreseeable future. We're reducing our reliance on GCP for some things, but I don't expect us to ever be AWS-only.
Hey Anurag! Thanks for answering our questions here. Just curious, why do most of the new PaaSs (including Render apparently) go for GCP and AWS, instead of a cheaper option like Hetzner? It seems everyday I read another blog post of how switching away from the Big 3 Clouds (or DO) results in massive savings for some company.
Hard to blame them. Google shutting down Domains, instead of rolling it into Cloud Domains, seems like a harbinger of what they’re planning for Cloud. Why offload it when they have a natural product to put it under, unless they’re also planning to shut down that product?
When I hear about a raise, especially this large, my initial thoughts are, "great! they must be doing something right", then I remember all those startups that raised mega rounds and ended up selling me as a customer down the river, either by the way of selling to someone like Google and writing an "amazing journey" blog post or just lock me in and jack up prices (I'm looking at you Customer.io).
That's why I always look for profitable companies that have built a business around solid foundations rather than raising external capital in round B, C, D all the way to Z.
If you want to see how those companies *really* were doing, look at their stock prices after they go public. Very few go up or even stay flat. Most of them crash right after the stock handcuffs come off and the investors have passed the bucket to the general public. I even have an anti-portfolio of public companies with investors who wouldn't shut up for a minute about how amazing they are doing, when they were private companies. All of them have done considerably worse than the market average or any basket of tech companies, even excluding the big five.
Have used Render for a few months to run some heavy worker scripts I wanted to place quickly -- it's a solid service but some of the UX is a bit rough and nearest DC to UK is Frankfurt although that turned out to not be an issue
I don't understand how any of these cloud wrapper companies will make their VC money back. Doesn't Vercel only have ~$25 million in ARR yet is somehow valued at $2.5 billion [0]?
If a company is using Vercel to start up, they're pretty good for what you get, but eventually AWS (which Vercel wraps) will win out in pricing, even counting hiring devops/cloud/backend engineers.
A way out for them would be to run their own infrastructure down the line which would allow them to pocket the (insane) margins the cloud providers currently make.
Looks like I can't see the tweet? Also, this lines up with Latka's figures as well, $24 million in 2022, and even with $50 million in 2023, that's still quite low for a $2.5 billion valuation, and that's if they don't raise further, thereby increasing their valuation.
Some things I like about Render vs Heroku since switching:
- Building the container from the Dockerfile in my repo seems to be a lot faster because it keeps all of the image layers cached.
- I can control the external IPs that have access to my postgres database. I usually keep them all off and only let the webapp access it internally, and then whitelist an IP as needed.
- Env groups are nice - if I want to start another service I can have it use all the same environment variables vs trying to keep them in sync myself.
A couple of years ago they had some multi-hour outages but things have been pretty stable since, a couple of the multi-minute outages were on me as I was upgrading the database or messing around with some configs. https://status.pinnacleclimb.com/
One thing I'd love in the future is some sort of postgres upgrade button. They handle minor version updates but to do a major version I need to put the app in maintenance mode, dump, restore, and swap to the new DB. It would be amazing if that could be automated with something like pglogical.
huge congrats to the Render team! especially raising a big round at a time when many VC insiders are saying things like the "series B and above market is closed".
I particularly remember the mad flurry by "cloud distro" startups to capitalize on Heroku essentially shutting down their free tier last year. It's an infrequent thing but I think having a playbook for how to respond when a major competitor churns up their own users on a silver platter is super helpful for future reference.
anything you'd call out as particularly helpful to you in the "post heroku wars" period? looking for specific principles i can extrapolate to other similar situations
no magic sauce; the most helpful thing was just staying up despite the massive influx of users and getting even better at fighting platform abuse. This may be very specific to what we do as a cloud provider.
i mean its also particularly churny. there's no "one standard way to do things" anymore. every framework/open source app launch now has to include instructions for like 4 different deploy targets because people can barely even standardize on helm charts much less how to orchestrate containers without k8s/docker compose (every newroku startup seems to want to reinvent the docker compose file, render included)
not complaining, innovation is good etc, but it doesnt come with out its costs to DX
Readit News
Funny enough I had big VC call me for due diligence (because I posted a lot on their forums and they spotted my name) and I happily told them how pissed I was with that change and how I went from vocal evangelist for the platform to looking for an alternative.
Still not as _terrible_ as Vercel which straight up doesn't deploy if the user isn't a paid org member lol. Now that is shameful as shit!
But this pricing change from Render was bad yeah. :(
We're looking for great people to join our team! See render.com/careers for all open roles.
I have _many_ tiny-usage projects, I'll say "10" but really many more. The thing is, even if you put their usage (resources, network, disk, etc) together it's still a "small" project equivalent, something that could run (capacity-wise) on a single VPS easily for $20-30/month. But the way Render is structured, even if I get 10 requests/month on each of them, I need to pay that per project. Which adds up quickly when you have many small projects as a hobby.
So I'm wondering if something like what Heroku had where there's a shared pool of 2000 credits that I can use among my projects as they get used by people, and when one of them becomes more popular then I can "upgrade" it to have a proper full-instance for each of its parts while keeping the rest of them in the pool.
If I wasn't so bad at devops, here the equivalent is I'd rent a mid-size VPS and have one folder/project, and I would be able to have them all hosted in that server with a lot of capacity to spare for $20-30/total. The way I have it now, if I wanted to host the same thing with Render, it'd be $250+/month only for my active projects. Heck, I think even my Raspberry Pi would be able to run all of these projects of mine properly (again, capacity-wise, I def love Render for its easiness of use).
I ultimately went with a large-ish droplet from DO with dokku installed. It’s been great, for the many small sites use case. It doesn’t scale, but that’s fine. When projects get traction the intention is to move them.
Which one has, and is now on AWS. But if we’d been on render I wouldn’t have moved it. I’d just spin up its own db and scale up the processes.
Second, have you guys had a chance to revise your hiring pipeline? When I spoke to your recruitment team a bit over a year ago they were asking for 8 straight hours on a zoom call as part of the software engineering interview. I had to drop out of the process because fitting that in was exceptionally difficult. (They did offer to split it into 5 and 3 but I still couldn't fit that in). They remarked they were looking to update that process a bit.
The above aside - I remember thinking everyone I spoke to there was pleasant and enjoyed working there. I was sad to have to drop out.
Best of luck moving forward.
I know this is my entitled Big Tech worker bias, but... surely there should be a lunch break?
I was very impressed by render when I was investigating hosting platforms, but if I'm being honest I got scared away by the postgres pricing(free for only 90 days: https://render.com/docs/free#free-postgresql-databases).
Question: After some more research, I noticed that even among competitors there aren't really any "always free" options(e.g. even AWS directly only offers 1 year free for RDS). What is it that makes free-tier postgres so uncompetitive in comparison to hosted compute? Is it because dedicated disk is more expensive than shared cores?
https://supabase.com/pricing (not affiliated, but a happy user)
I'd like to give it a shot but I often don't get past HR because I'm in (theoretical) CS and lack "real world experience". Which is lame because the stuff I've done from my trench is amazing (source: trust me, lol).
Anyway, I'm really good on the backend, creating high performance REST APIs, documenting them, etc... and know a great deal of postgres.
I also have a good intuition for what's going to stick or not in the field. I was doing AWS ~15 years ago, back when you had to use an ugly SOAP interface to get things done, but I saw the potential and created the first automated cluster of bioinformatic tools in there. Right now, I'm betting on Rust, WASM, V8 isolates, and everything that has to do w/ lightweight function execution.
If that sounds good for render's mission, I'll be glad to take part in it.
1. Company starts to do X, makes it a wonderful experience
2. Users love it, it's successful
3. Company stop treating users well
4. New company steps into make X a wonderful experience again, with the advantage that the market is pre-validated.
How much of an advantage is it to know that users want what you're creating? Do you have any tips, like ways to reassure people you won't follow the same dynamic as the original company?
I wonder what the ideal time is in a company's arc to start their eventual replacement. Render just raised a series B which probably means they're still many years away from step 3, so it's probably too early. But maybe when they're raising a series C or D, it's time to start thinking about making their replacement.
I feel like Stripe is entering that territory right now. Not that they're worse than alternatives, but they no longer have that "wonderful experience" magic because they've started to turn on the maximize shareholder value engine.
With Render, I'd rather show than tell. As an interesting data point, we've now existed as an independent entity longer than Heroku did, and we're truly just getting started.
> 100% employer-paid medical coverage and 99% employer-paid dental and vision coverage for you and all your dependents. FSAs available too.
Good luck
That's why I always look for profitable companies that have built a business around solid foundations rather than raising external capital in round B, C, D all the way to Z.
If you want to see how those companies *really* were doing, look at their stock prices after they go public. Very few go up or even stay flat. Most of them crash right after the stock handcuffs come off and the investors have passed the bucket to the general public. I even have an anti-portfolio of public companies with investors who wouldn't shut up for a minute about how amazing they are doing, when they were private companies. All of them have done considerably worse than the market average or any basket of tech companies, even excluding the big five.
There is a lot they are getting right
If a company is using Vercel to start up, they're pretty good for what you get, but eventually AWS (which Vercel wraps) will win out in pricing, even counting hiring devops/cloud/backend engineers.
[0] https://getlatka.com/companies/vercel
rauchg disclosed their revenue trends recently https://twitter.com/swyx/status/1667995063216148480 i learned later that this interview was done in nov 2022
Some things I like about Render vs Heroku since switching:
A couple of years ago they had some multi-hour outages but things have been pretty stable since, a couple of the multi-minute outages were on me as I was upgrading the database or messing around with some configs. https://status.pinnacleclimb.com/One thing I'd love in the future is some sort of postgres upgrade button. They handle minor version updates but to do a major version I need to put the app in maintenance mode, dump, restore, and swap to the new DB. It would be amazing if that could be automated with something like pglogical.
I particularly remember the mad flurry by "cloud distro" startups to capitalize on Heroku essentially shutting down their free tier last year. It's an infrequent thing but I think having a playbook for how to respond when a major competitor churns up their own users on a silver platter is super helpful for future reference.
anything you'd call out as particularly helpful to you in the "post heroku wars" period? looking for specific principles i can extrapolate to other similar situations
not complaining, innovation is good etc, but it doesnt come with out its costs to DX