checkpoint-writer

You can resume a multi-phase pipeline by triggering a checkpoint at the end of each phase. This saves the current pipeline state to disk, allowing you to seamlessly resume the workflow from the exact interruption point after a failure or restart.

Checkpointing in a DevOps pipeline is the process of saving the current phase state to disk to enable reliable recovery. You need it during long-running pipelines where interruptions or failures could cause significant loss of state and require a full restart.

Yes, this state-saving approach integrates with existing git workflows by optionally recording a git-based anchor commit. This provides provenance and auditability for the pipeline state recovery point saved to disk.

To create a recoverable checkpoint, trigger the save mechanism at the end-of-phase points in your multi-phase workflow. It implements filesystem writes to a specific checkpoints directory and guards against first-phase start errors before session termination.

Yes, you can restart a pipeline from a checkpoint without losing state. The mechanism saves a deterministic recovery point at the end of each phase, allowing you to reconstruct the pipeline state and resume seamlessly across restarts.