8 ms·

Classic Salesforce. The exact same thing happened with our org and Heroku. Zero empathy, just pony up or we trash your company.

jonplackett 1y ago
Yeah they fucked Heroku hard. I used to love Heroku. Can’t imagine there’s many people still left using it now.
- mrroryflint 1y ago
  Hundreds of thousands, in fact. But I bet it’s a downward trend with no hope of a turnaround.
- jorisboris 1y ago
  I still have old personal projects on there Its inertia, its just not a priority to move them over
- wereHamster 1y ago
  We just managed to shut down our last Heroku service a week ago. Good riddance.
  - gregsadetsky 1y ago
    what did you migrate to?
    - TuringNYC 1y ago
      Not OP, but we originally migrated to Digital Ocean. But now, all complex stuff is on AWS and dual-stack stuff is on Replit.
      - gregsadetsky 1y ago
        and what did you use to manage git push deployments, setting env vars to replicate the heroku features? also, you used replit for the frontend deployment? or frontend and some backend? thanks - just super interesting as I'm in the space and feedback/real cases are really useful
        TuringNYC 1y ago
        >> and what did you use to manage git push deployments, setting env vars to replicate the heroku features? Yes Digital Ocean did all this, they were very feature-close to Heroku. We have over time migrated everything stable/prod to AWS just because AWS has more products and hence you have everything in one place inside a VPC (e.g. vector db) For Replit, i'd use it for anything I can in early-stages. It helps to prototype ideas you are testing. You can iterate rapidly. For PROD we'd centralize onto AWS given the ecosystem.
        gregsadetsky 1y ago
        cheers, I really appreciate your answers and last q :-) re AWS - once you moved there, did you use something like elasticbean or app runner? or did you roll your own CI/CD/logging/scaling...?
        TuringNYC 1y ago
        > and last q :-) re AWS - once you moved there, did you use something like elasticbean or app runner? or did you roll your own CI/CD/logging/scaling...? We started with Lambdas because you can split work across people and keep dependencies to a minimum. Once your team gels and your product stabilizes, it is helpful to Dockerize it and go ECS, that is what we did. Some teams in the past used EKS but IMHO it required too much knowledge for the team to maintain, hence we've stuck with ECS. All CI/CD via Github --> ECS. This is a very standard pipeline and works well locally for development also. ECS does the scaling quite well, and provides a natural path to EKS when you need the scale bigtime. For logging, if I could choose I'd go Datadog but often you go with whatever the budget solution is.
    - wereHamster 1y ago
      We got rid of all Rails apps (that needed a backend). We've moved our Postgres databases to Neon, and run our docker containers on Google Cloud Run (these are containers that don't need to run 24/7, we're paying just a few cents each month, also cold starts are much faster and more reliable than on Heroku).
- Mo3 1y ago
  > Yeah they fucked Heroku hard Surprisingly not as much as I'd thought when they took it over. They just never adjusted pricing to remain competitive. The experience is still some of the best you can get for RoR apps. But nobody in their right mind deploying a new application today would look at their insane 10 year old dyno pricing and be like - yup - reasonable
  - jrochkind1 1y ago
    in fact, if you actually look at the historical timeline, many of the things we think of as core to developer experience only were released after salesforce acquisition. I think even multiple buildpacks at once only came a couple years after acquisition. Possibly they were in the pipeline before acquisition, sure. But I'd agree, heroku is still a better DX than almost any competitors, although it's features and pricing have really stagnated. So better DX as long as you don't need any features it doens't have. But it hasn't really been 'ruined' in any way, it just started appearing frozen in amber some years ago. The new 'fir' platform is promissing, before that I didn't really know that any actual development was taking place in heroku, but it's a big move, modernizing things and setting the stage for more. Including slightly improved resource-to-pricing options. We'll see if it all works out...
joshgachnang 1y ago
Slack just did the same to us for our company Slack. We have to have the HIPAA compliant Enterprise version, price going up 40% next year. Looks like we'll migrate, especially because compliance has a bunch of annoying caveats.
- Corrado 1y ago
  We're in the same boat; HIPAA compliant Enterprise license. Slack came to us with a 2 day notice; pay more now or pay a lot more later. We asked if we could reduce the number of users and they said no, if you change anything then you have to take the new pricing for double the current price. The whole thing was super sleazy. We told them that we were moving to MS Teams (arrrgghhh!) and they said "Bye!".
  - joshgachnang 1y ago
    Ohh woooow, they at least gave us like 9 months (until our next contract renewal). Right now the plan is to move to Google Chat (already a Google shop) and an internally build chat in our EHR for some of the patient-focused things.

jonplackett 1y ago

Yeah they fucked Heroku hard. I used to love Heroku. Can’t imagine there’s many people still left using it now.

mrroryflint 1y ago
Hundreds of thousands, in fact. But I bet it’s a downward trend with no hope of a turnaround.
jorisboris 1y ago
I still have old personal projects on there Its inertia, its just not a priority to move them over
wereHamster 1y ago
We just managed to shut down our last Heroku service a week ago. Good riddance.
- gregsadetsky 1y ago
  what did you migrate to?
  - TuringNYC 1y ago
    Not OP, but we originally migrated to Digital Ocean. But now, all complex stuff is on AWS and dual-stack stuff is on Replit.
    - gregsadetsky 1y ago
      and what did you use to manage git push deployments, setting env vars to replicate the heroku features? also, you used replit for the frontend deployment? or frontend and some backend? thanks - just super interesting as I'm in the space and feedback/real cases are really useful
      - TuringNYC 1y ago
        >> and what did you use to manage git push deployments, setting env vars to replicate the heroku features? Yes Digital Ocean did all this, they were very feature-close to Heroku. We have over time migrated everything stable/prod to AWS just because AWS has more products and hence you have everything in one place inside a VPC (e.g. vector db) For Replit, i'd use it for anything I can in early-stages. It helps to prototype ideas you are testing. You can iterate rapidly. For PROD we'd centralize onto AWS given the ecosystem.
        gregsadetsky 1y ago
        cheers, I really appreciate your answers and last q :-) re AWS - once you moved there, did you use something like elasticbean or app runner? or did you roll your own CI/CD/logging/scaling...?
        TuringNYC 1y ago
        > and last q :-) re AWS - once you moved there, did you use something like elasticbean or app runner? or did you roll your own CI/CD/logging/scaling...? We started with Lambdas because you can split work across people and keep dependencies to a minimum. Once your team gels and your product stabilizes, it is helpful to Dockerize it and go ECS, that is what we did. Some teams in the past used EKS but IMHO it required too much knowledge for the team to maintain, hence we've stuck with ECS. All CI/CD via Github --> ECS. This is a very standard pipeline and works well locally for development also. ECS does the scaling quite well, and provides a natural path to EKS when you need the scale bigtime. For logging, if I could choose I'd go Datadog but often you go with whatever the budget solution is.
  - wereHamster 1y ago
    We got rid of all Rails apps (that needed a backend). We've moved our Postgres databases to Neon, and run our docker containers on Google Cloud Run (these are containers that don't need to run 24/7, we're paying just a few cents each month, also cold starts are much faster and more reliable than on Heroku).
Mo3 1y ago
> Yeah they fucked Heroku hard Surprisingly not as much as I'd thought when they took it over. They just never adjusted pricing to remain competitive. The experience is still some of the best you can get for RoR apps. But nobody in their right mind deploying a new application today would look at their insane 10 year old dyno pricing and be like - yup - reasonable
- jrochkind1 1y ago
  in fact, if you actually look at the historical timeline, many of the things we think of as core to developer experience only were released after salesforce acquisition. I think even multiple buildpacks at once only came a couple years after acquisition. Possibly they were in the pipeline before acquisition, sure. But I'd agree, heroku is still a better DX than almost any competitors, although it's features and pricing have really stagnated. So better DX as long as you don't need any features it doens't have. But it hasn't really been 'ruined' in any way, it just started appearing frozen in amber some years ago. The new 'fir' platform is promissing, before that I didn't really know that any actual development was taking place in heroku, but it's a big move, modernizing things and setting the stage for more. Including slightly improved resource-to-pricing options. We'll see if it all works out...

joshgachnang 1y ago

Slack just did the same to us for our company Slack. We have to have the HIPAA compliant Enterprise version, price going up 40% next year. Looks like we'll migrate, especially because compliance has a bunch of annoying caveats.

Corrado 1y ago
We're in the same boat; HIPAA compliant Enterprise license. Slack came to us with a 2 day notice; pay more now or pay a lot more later. We asked if we could reduce the number of users and they said no, if you change anything then you have to take the new pricing for double the current price. The whole thing was super sleazy. We told them that we were moving to MS Teams (arrrgghhh!) and they said "Bye!".
- joshgachnang 1y ago
  Ohh woooow, they at least gave us like 9 months (until our next contract renewal). Right now the plan is to move to Google Chat (already a Google shop) and an internally build chat in our EHR for some of the patient-focused things.