Archiving a task deletes its row from the task table but leaves its events, transitions, step executions and model usage behind, with no parent row to join and nothing that prunes them. On the live instance 22 archived tasks left 1,140 events, 130 transitions, 420 step executions and 231 model usage rows unreachable.
Wanted: decide and implement the retention rule — prune a task's child rows with the archive after an audit window, or keep them deliberately (and make them visible through the audit feed) — and say which in the archived-task record.
Done when: archiving either removes or accounts for every child row, and a query can tell what happened to an archived task.
Archiving a task deletes its row from the task table but leaves its events, transitions, step executions and model usage behind, with no parent row to join and nothing that prunes them. On the live instance 22 archived tasks left 1,140 events, 130 transitions, 420 step executions and 231 model usage rows unreachable.
Wanted: decide and implement the retention rule — prune a task's child rows with the archive after an audit window, or keep them deliberately (and make them visible through the audit feed) — and say which in the archived-task record.
Done when: archiving either removes or accounts for every child row, and a query can tell what happened to an archived task.
Already resolved by c155bb24 on main: archiving deliberately keeps a task's child rows (transitions, step executions, model usage, tool calls, dispatch ledger), and the archive event, which audit retention never prunes and the audit feed shows, records that rule in its data (retention: kept, the child sets, events governed by audit_retention). That satisfies the issue's 'keep deliberately and say which in the archived-task record' branch. No further code change; nothing was run for this close beyond confirming the commit is an ancestor of origin/main.
Already resolved by c155bb24 on main: archiving deliberately keeps a task's child rows (transitions, step executions, model usage, tool calls, dispatch ledger), and the archive event, which audit retention never prunes and the audit feed shows, records that rule in its data (retention: kept, the child sets, events governed by audit_retention). That satisfies the issue's 'keep deliberately and say which in the archived-task record' branch. No further code change; nothing was run for this close beyond confirming the commit is an ancestor of origin/main.
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Archiving a task deletes its row from the task table but leaves its events, transitions, step executions and model usage behind, with no parent row to join and nothing that prunes them. On the live instance 22 archived tasks left 1,140 events, 130 transitions, 420 step executions and 231 model usage rows unreachable.
Wanted: decide and implement the retention rule — prune a task's child rows with the archive after an audit window, or keep them deliberately (and make them visible through the audit feed) — and say which in the archived-task record.
Done when: archiving either removes or accounts for every child row, and a query can tell what happened to an archived task.
Already resolved by
c155bb24on main: archiving deliberately keeps a task's child rows (transitions, step executions, model usage, tool calls, dispatch ledger), and the archive event, which audit retention never prunes and the audit feed shows, records that rule in its data (retention: kept, the child sets, events governed by audit_retention). That satisfies the issue's 'keep deliberately and say which in the archived-task record' branch. No further code change; nothing was run for this close beyond confirming the commit is an ancestor of origin/main.