I have a cloud application in which my worker process that is using ClaudeSDKClient (the Python API) could crash. I need to resume my conversation, but the worker's file system is essentially inaccessible.
How might I resume the conversation in another worker? From reading anthropics/claude-code#5135, I suspect the most viable workaround is to persist my own independent conversation logs and then munge them into a jsonl file in a spoofed directory on my new worker instance.
Is that true? If there's not a better solution, is one planned?
I have a cloud application in which my worker process that is using
ClaudeSDKClient(the Python API) could crash. I need to resume my conversation, but the worker's file system is essentially inaccessible.How might I resume the conversation in another worker? From reading anthropics/claude-code#5135, I suspect the most viable workaround is to persist my own independent conversation logs and then munge them into a
jsonlfile in a spoofed directory on my new worker instance.Is that true? If there's not a better solution, is one planned?