How do you safely experiment with Convex schema changes?
❔Advice
I am asking for your advice, so since this is not a feature request, or bug report, i believe the version im in or other details are not that important, but let's say latest version + javascript(nextjs).
I love working with Convex — amazing DX — but I always hit the same problem.
We share one dev Convex deployment.
Whenever I want to experiment with schema changes (rename tables, redesign models, test v2 ideas), I have to:
1- delete existing records
2- create temporary tables
3- rename things back and forth
4- clean up the DB manually
This quickly becomes messy and confusing.
What I want:
- Open a new Git branch
- Have an isolated Convex environment for that branch (separate DB)
- Freely break / change schemas
- Delete the branch → everything is gone
Because we have to work on different branches / different tasks without finishing the other.
I checked Preview Deployments, but I’m not sure they solve this use case.
Question:
What’s the recommended workflow in Convex for experimenting with schema changes without constantly deleting shared dev data?
Do people usually create separate dev deployments per developer / feature, or is there a better approach?
Thx everyone.
I love working with Convex — amazing DX — but I always hit the same problem.
We share one dev Convex deployment.
Whenever I want to experiment with schema changes (rename tables, redesign models, test v2 ideas), I have to:
1- delete existing records
2- create temporary tables
3- rename things back and forth
4- clean up the DB manually
This quickly becomes messy and confusing.
What I want:
- Open a new Git branch
- Have an isolated Convex environment for that branch (separate DB)
- Freely break / change schemas
- Delete the branch → everything is gone
Because we have to work on different branches / different tasks without finishing the other.
I checked Preview Deployments, but I’m not sure they solve this use case.
Question:
What’s the recommended workflow in Convex for experimenting with schema changes without constantly deleting shared dev data?
Do people usually create separate dev deployments per developer / feature, or is there a better approach?
Thx everyone.
