As someone who has used and enjoyed Heroku off-and-on since 2010, I was rattled by the phrasing of the announcement.
Reading comments about people's challenges and displeasures with Heroku over the years, they have almost never resonated with me. When the complaints were contextualized, I certainly understood them, but they have not been applicable to my needs and experiences.
My current team at work had a meeting about the announcement, and decided to spend gradual time over the next year exploring how we would migrate off Heroku if we must, and running tests of our own alternative infra in pursuit of that. It is also our desire not to need to! Our first-pass assessment of such a migration is that it would (1) be time-consuming at the expensive of other work, (2) be more expensive (in engineering time) than we presently spend, and (3) likely result in worse DX than what Heroku provides.
We definitely don't want to leave, but we also know the professional choice is to be prepared to do so within the next year or two. We would not have had that conversation at all if the announcement had not been so strange. If I have any feedback for the leadership at Salesforce, it would be that: communicate better, because you are pushing otherwise-satisfied customers away.
If it works fine and cash flows then sales force would sell it before shutting it down. PE would love to acquire and milk it (and you).
Honestly though it isn’t that hard to go k8s anymore and self host with Argo etc. You can use ChatGPT to figure it all out. Just go bare minimum commodity VPS and use agnostic code as infrastructure like terraform. Then you can just win from the race to the bottom of cost
As someone who moved from Heroku -> App Engine -> Cloud Run, I think you'll appreciate the modern alternatives more. If you remove the cost factor, the development experience within GCP is far superior. Not to mention the security features are great as well.