Thoughts on Render.com
medium.com
medium.com
The actual services worked great, but their pricing changes at short notice left me in an awkward spot that made me not trust Render going forwards.
tl;dr is that they introduced a charge of an additional $20/mo per person on top of compute. Even if it's just to maintain billing info.
This may be trivial if you're a business spending hundreds or thousands on compute every month, but for bootstrappers, hobbyists and freelancers this is horrible. A $7 instance easily becomes $67/m if a couple of other people need access.
I had just moved a bunch of services over from Heroku when this price change was announced, and it was a very sore experience. Poorly communicated with very short notice.
Jumped ship again to DigitalOcean, which has a near identical product but without the wonky pricing.
More discussion: https://community.render.com/t/pricing-update-nov-2022-reque...
The team charge assigns a nominal value to the productivity boost Render gives to teams, without increasing compute prices (which stayed the same even with the change). Existing single-seat teams also got a 6-month grace period for team fees.
While our change means we're not the cheapest option on the market, we believe the DX and features we have (and will continue to add) makes Render an incredibly good deal, both now and over the long term.
I'd be really interested in hearing why your reasoning for seat-based pricing, as I struggle to see this from a user perspective, and from the perspective as someone who runs a SaaS. Paying a $20/month premium per user just for the privilege of paying you for computing seems like a strange proposal.
Now, I'm currently on heroku, and to be honest, the pricing of heroku is not a huge problem for me, although did seem a bit steep. (With a small/medium sized project for my small non-profit employer, I pay heroku somewhere between 8-10K a year -- kind of a lot for us, but less than sysadmin/devops, and peanuts for heroku).
Render used to seem _incredibly_ lower-priced than heroku, which led some people to suggest that heroku was incredibly over-priced... but perhaps this pricing change shows that in fact render was kind of under-priced. It's hard to say exactly what my spend on render would be if I switched, as the metrics charged are different, and I haven't done the work to estimate.
But I'd guess it's still a bit cheaper -- and honestly even if it were about the same, that would be fine... we're not looking to get off heroku because of price mostly, but because we don't trust heroku's future.
I really don't love the per-developer pricing or the non-trivial bandwidth pricing though. I don't like being incentivized to have fewer developers with access to render. And for a small organization, having predictable pricing that doesn't go up if our bandwidth goes up (not entirely under our control) is huge, and something that heroku gives us... although I realize there are hardly any services that can give you unmetered bandwidth these days, it may just not be feasible.
I would _rather_ have increased compute prices, that are more predictable and stable costs for us, instead of a per-developer or bandwidth increase.
I might be missing something, but App platform provides the same features and more for a cheaper price.
But otherwise there's very little difference. DOAP ends up a lot cheaper overall too.