Hello!

• Vote on this issue by adding a 👍 reaction
• If you want to implement this feature, comment to let us know (we'll work with you on design, scheduling, etc.)

Issue details

We want to use Pulumi cloud as much as possible. But there are continuous internal/edge deployments that only have occasional internet connections. We want to make sure that internet outage will not block internal deployments. We want to sync the state back to Pulumi cloud when internet connection is back, so that people would primarily use the cloud interface knowing there might be delayed state updates.

Affected area/feature

It'd be great if pulumi can support primary/backup state backends such that pulumi operations go through the primary backend (e.g., cloud or local s3) and replicate the state to a backup backend. If the primary backend is not available, failover to the backup backend and deployments are not blocked. When the primary comes back, the state is sync'ed back to the primary.

It's possible to jerry rig something with export/import on our end, but it's hard to do clean/raceless switches between primary and backup without modifying pulumi.

If anybody has cleaner workarounds, we'd like to hear them as well. Thanks in advance!