Skip to content

Reconcile late durable workflow outcomes #324

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
witbrock merged 1 commit into from
Jul 28, 2026
Merged

Reconcile late durable workflow outcomes #324

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
67 changes: 66 additions & 1 deletion src/backend/integrations/internal_mcp/catalogue.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -11437,6 +11437,61 @@ def _summarise_turn_execution_tool_invocations(
entry["tool"] = tool_name
if status_text is not None:
entry["status"] = status_text
for field_name in (
"call_id",
"execution_id",
"effect_id",
"effect_status",
"error_code",
"mutation_outcome",
"outcome_finality",
"capability_kind",
"execution_method",
"represented_workflow_id",
"workflow_id",
"instance_id",
"durable_submission_status",
"final_status",
):
value = raw_entry.get(field_name)
if isinstance(value, str) and value.strip():
entry[field_name] = value.strip()
evidence = raw_entry.get("evidence")
if isinstance(evidence, Mapping):
evidence_provenance = (
evidence.get("provenance")
if isinstance(evidence.get("provenance"), Mapping)
else {}
)
for field_name in (
"execution_id",
"effect_id",
"effect_status",
"error_code",
"mutation_outcome",
"outcome_finality",
"workflow_id",
"instance_id",
"durable_submission_status",
"final_status",
):
if field_name in entry:
continue
value = evidence.get(field_name)
if isinstance(value, str) and value.strip():
entry[field_name] = value.strip()
for field_name in (
"capability_kind",
"execution_method",
"represented_workflow_id",
):
if field_name in entry:
continue
value = evidence_provenance.get(field_name)
if isinstance(value, str) and value.strip():
entry[field_name] = value.strip()
if isinstance(raw_entry.get("changed"), bool):
entry["changed"] = raw_entry.get("changed")
summary.append(entry)
if len(summary) >= max_items:
break
Expand Down Expand Up @@ -11635,6 +11690,10 @@ def _extract_turn_execution_tool_invocation_summary(
if isinstance(existing_summary, list):
return _summarise_turn_execution_tool_invocations(existing_summary)

direct_invocations = item.get("tool_invocations")
if isinstance(direct_invocations, list):
return _summarise_turn_execution_tool_invocations(direct_invocations)

execution_raw = item.get("execution")
execution = execution_raw if isinstance(execution_raw, Mapping) else {}
return _summarise_turn_execution_tool_invocations(execution.get("tool_invocations"))
Expand Down Expand Up @@ -23200,7 +23259,13 @@ def _rag_get_item(**kwargs):
execution_summary: Mapping[str, Any] = (
execution_summary_raw if isinstance(execution_summary_raw, Mapping) else {}
)
tool_invocations = execution_payload.get("tool_invocations")
direct_tool_invocations = doc.get("tool_invocations")
tool_invocations = (
direct_tool_invocations
if isinstance(direct_tool_invocations, list)
and direct_tool_invocations
else execution_payload.get("tool_invocations")
)
discovery_payload_raw = workflow_routing_diagnostics.get("discovery")
discovery_payload: Mapping[str, Any] = (
discovery_payload_raw if isinstance(discovery_payload_raw, Mapping) else {}
Expand Down
143 changes: 133 additions & 10 deletions src/backend/services/turn_execution_diagnostics_service.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -524,6 +524,119 @@ def _normalise_tool_history(tool_history: Any) -> list[dict[str, Any]]:
return []


def _summarise_tool_invocation_history(
tool_invocations: Any,
) -> list[dict[str, Any]]:
"""Build a bounded diagnostic history from ordinary-turn invocations."""

if not isinstance(tool_invocations, list):
return []
summary: list[dict[str, Any]] = []
for raw_entry in tool_invocations:
if not isinstance(raw_entry, Mapping):
continue
tool_name = (
_safe_str(raw_entry.get("tool"))
or _safe_str(raw_entry.get("tool_name"))
or _safe_str(raw_entry.get("method"))
)
if not tool_name:
continue
entry: dict[str, Any] = {"tool": tool_name}
for key in (
"status",
"call_id",
"effect_id",
"effect_status",
"error_code",
"execution_id",
"capability_kind",
"execution_method",
"represented_workflow_id",
"workflow_id",
"instance_id",
"durable_submission_status",
"mutation_outcome",
"outcome_finality",
"final_status",
"result_summary",
):
value = raw_entry.get(key)
if isinstance(value, str) and value.strip():
entry[key] = value.strip()
if isinstance(raw_entry.get("changed"), bool):
entry["changed"] = raw_entry.get("changed")
transport = raw_entry.get("transport")
if isinstance(transport, Mapping):
entry["transport"] = {
key: transport.get(key)
for key in (
"schema_version",
"execution_id",
"outcome",
"duration_ms",
"timeout_sec",
"advisory_timeout_sec",
"queue_duration_ms",
"handler_duration_ms",
"handler_elapsed_ms",
"timeout_phase",
"late_result_policy",
)
if key in transport
}
if "execution_id" not in entry:
execution_id = _safe_str(transport.get("execution_id"))
if execution_id:
entry["execution_id"] = execution_id
evidence = raw_entry.get("evidence")
if isinstance(evidence, Mapping):
for key in (
"effect_status",
"error_code",
"instance_id",
"workflow_id",
"durable_submission_status",
"mutation_outcome",
"outcome_finality",
"final_status",
):
if key in entry:
continue
value = evidence.get(key)
if isinstance(value, str) and value.strip():
entry[key] = value.strip()
summary.append(entry)
return summary


def _fallback_tool_invocation_history(
*,
execution: Mapping[str, Any] | None,
turn_record: Mapping[str, Any] | None,
llm_debug: Mapping[str, Any] | None,
) -> list[dict[str, Any]]:
candidates = (
execution.get("tool_invocations")
if isinstance(execution, Mapping)
else None,
turn_record.get("tool_invocations")
if isinstance(turn_record, Mapping)
else None,
llm_debug.get("tool_invocations")
if isinstance(llm_debug, Mapping)
else None,
llm_debug.get("invocations")
if isinstance(llm_debug, Mapping)
else None,
)
for candidate in candidates:
summary = _summarise_tool_invocation_history(candidate)
if summary:
return summary
return []


def _derive_tool_counts(tool_history: Sequence[Mapping[str, Any]]) -> dict[str, int]:
total = 0
success = 0
Expand All @@ -534,7 +647,7 @@ def _derive_tool_counts(tool_history: Sequence[Mapping[str, Any]]) -> dict[str,
continue
total += 1
status = (_safe_str(entry.get("status")) or "").lower()
if status in {"completed", "complete", "success", "succeeded"}:
if status in {"ok", "completed", "complete", "success", "succeeded"}:
success += 1
ended += 1
elif status in {
Expand All @@ -544,6 +657,11 @@ def _derive_tool_counts(tool_history: Sequence[Mapping[str, Any]]) -> dict[str,
"blocked",
"cancelled",
"canceled",
"partial",
"indeterminate",
"not_started",
"timeout",
"timed_out",
}:
failure += 1
ended += 1
Expand Down Expand Up @@ -1087,8 +1205,10 @@ def _build_fallback_turn_execution_diagnostics(
)
)

tool_history = _normalise_tool_history(
execution.get("tool_invocations") if execution else None
tool_history = _fallback_tool_invocation_history(
execution=execution,
turn_record=turn_record,
llm_debug=llm_debug_mapping,
)
tool_counts = _derive_tool_counts(tool_history)

Expand Down Expand Up @@ -1344,16 +1464,19 @@ def _normalise_embedded_diagnostics_payload(
)

tool_history = _normalise_tool_history(payload.get("tool_history"))
tool_invocations = execution.get("tool_invocations") if execution else None
if not tool_history and isinstance(tool_invocations, list):
tool_history = _normalise_tool_history(tool_invocations)
payload["tool_history"] = tool_history
else:
payload["tool_history"] = tool_history
hydrated_tool_history = False
if not tool_history:
tool_history = _fallback_tool_invocation_history(
execution=execution,
turn_record=turn_record,
llm_debug=llm_debug_mapping,
)
hydrated_tool_history = bool(tool_history)
payload["tool_history"] = tool_history

tool_counts = _derive_tool_counts(tool_history)
for key, value in tool_counts.items():
if not isinstance(payload.get(key), int):
if hydrated_tool_history or not isinstance(payload.get(key), int):
payload[key] = value

_attach_minimal_turn_timing_trace(
Expand Down
Loading
Loading