How to support langgraph's short term memory?

[shriyanskapoor]

Hi,

is there a way to support checkpointing in microservice like agent architecture where individual agents are deployed on different endpoints?

Also on the same line, how to pause and resume execution at a specific agent/step?

For details have a look a much more descriptive question I asked on langgraphs github https://github.com/langchain-ai/langgraph/discussions/5036

[insanesac]

not sure about the first one, but you could try setting up tasks for the second question. i saw that in one of their examples they mentioned using task to cancel a run.