Reuse valid workspace checkpoints on request

This commit is contained in:
2026-07-08 03:02:50 +00:00
parent 1d3a444df8
commit ae9c2e1d5e
14 changed files with 1406 additions and 290 deletions

View File

@@ -74,6 +74,8 @@ Implemented diagnostics artifacts:
- `resolved-references.json`: resolved reference provenance, including target
stage, lane ID when present, origin, digest, media type, byte size, and
binding source, without reference content.
- `checkpoint-events.json`: checkpoint steps that were reused or executed
during an explicit resume invocation.
- `run-manifest.json`: the same run manifest written to durable output when it
is available, including top-level module metadata when present.
- `warnings.json`: warning list.
@@ -94,14 +96,21 @@ write checkpoints under:
```
Each workflow step owns its own manifest and payload files. There is no
root-level checkpoint summary. Current runs write checkpoints for inspection and
future recovery support only; the CLI does not read checkpoints or skip work.
root-level checkpoint summary. Ordinary `notarius run` invocations execute the
pipeline normally and refresh checkpoints. `notarius run --resume` reuses valid
checkpoints and executes any missing, invalid, or incompatible step normally.
Checkpoint payloads preserve byte content with base64 envelopes, media type,
metadata, warnings, and content digests where applicable. Checkpoints do not
include raw prompts, raw reference contents, raw LLM request payloads, or debug
traces.
A checkpoint is reused only when its workspace schema version, checkpoint
identity digest, step status, dependency fingerprints, payload files, and
payload digests match the current invocation. Changes to input bytes, resolved
pipeline digest, selected lanes, runtime LLM profile override, or materialized
reference digests invalidate reuse.
## Retention
Diagnostics retention is configured with `workspace.diagnostics.retention`,
@@ -156,8 +165,8 @@ directories unless they are part of your own operational policy.
## Operational Limits
Checkpoint writing does not provide resume execution yet. Re-run `notarius run`
after fixing the cause of a failed run.
If `--resume` cannot reuse a checkpoint, Notarius executes that step and writes
a fresh checkpoint when checkpointing is enabled.
Provider retries and timeouts are handled by Scriptorium according to the
selected execution profile. Pipeline module retries are controlled by module