Is it possible to save the state of a workflow instance, and restore its execution from the saved state? Possible use cases include 1) resuming the execution of a workflow from an intermediate stage for testing or postmortem analysis, 2) migrating workload from one zeebe cluster to another, i.e. upgrade to a backward incompatible version of zeebe, etc.
Hi @liuyu, it’s not possible to do this at the moment.
You can export the stream associated with a workflow, and reconstruct its state at any point via projection.
However, you can’t “hydrate” a workflow in the engine from that.
Guys, what would be the recommended strategy to restart a failed workflow from a certain state? We would not want to restart from the start.
Raise an incident, then use Operate to resolve the incident and resume the workflow. Watch the demo in this webinar to see it happening: https://camunda.com/learn/webinars/zeebe-mid-year-update-july-2019