Checkpointing

As explained in Building Reliable AI Systems

The practice of saving execution state so a process can be paused and later resumed.

Checkpointing saves agent execution state so work can pause and resume. In a harness–sandbox setup, the harness can restore state after a sandbox crash and continue from the last checkpoint. LangGraph provides checkpointing as part of its state management for production agents. section 6.2.2 section 7.9.1

Figure 7.4 The harness owns the agent loop, memory, and LLM communication. The sandbox executes code in isolation with no direct model access. Tasks flow right; results flow left. If the sandbox crashes, the harness recovers.

Shows the harness and sandbox roles, including recovery when the sandbox crashes. section 7.9.1

How Building Reliable AI Systems covers it

How deep it goes

Connects checkpointing to pause, resume, and crash recovery, but does not detail checkpoint formats or save-and-restore procedures.

Examples

Uses a sandbox crash as the recovery case: restore harness state and continue from the last checkpoint.

What you can do after

Understand how an agent harness can recover execution after a sandbox crash using the last checkpoint.

How it connects

Links checkpointing with agent state management, harnesses, and sandboxed execution.

liveBookmanning.com

Read Building Reliable AI Systems on liveBook Get the book at manning.com

Also explained in

• 

  liveBookmanning.com
  Architecting an Apache Iceberg LakehouseA checkpoint is persisted progress information from which an incremental data-processing workflow can resume after interruption.
• 

  liveBookmanning.com
  Kafka for ArchitectsCheckpointing is the periodic capture and persistence of streaming application state so it can be restored after failure.
• 

  liveBookmanning.com
  ScyllaDB in ActionCheckpointing records completed progress so that a process can pause and resume from an earlier point without restarting.
• 

  liveBookmanning.com
  AI Agents and Applications
• 

  liveBookmanning.com
  MongoDB 8.0 in Action, Third EditionA stream processing checkpoint is a persisted record of processing progress and state that allows a stream processor to resume consistently after interruption.

Compare how 6 books explain checkpointing →

Related concepts

• Flinkused by · Apache Flink is an open-source distributed processing framework for stateful computation over bounded and unbounded data streams.

This explainer was generated by Manning's knowledge graph from the text of Building Reliable AI Systems. The section links point to the parts of the book it draws on.