增强(08-23-isolated-backup-restore-drill): 固化八阶段RTO合同

2026-08-25 20:57:15 +08:00
@@ -1330,6 +1330,13 @@ validate_copied_batch(private/batch, source_metadata)
 - Exit-code fields are non-negative integers or `NOT_RUN`;
  `pg_restore_error_count` additionally permits `UNKNOWN` when no safe numeric
  summary is available.
+- Required data-restore timing fields:
+  `rto_batch_copy_seconds`, `rto_postgres_container_start_seconds`,
+  `rto_dump_list_seconds`, `rto_target_confirmation_seconds`,
+  `rto_restore_apply_seconds`, `rto_database_verify_seconds`,
+  `rto_storage_attachment_verify_seconds`, `rto_final_integrity_seconds` and
+  `data_restore_rto_seconds`. A measured value is a non-negative integer;
+  an unstarted stage is `NOT_RUN`.
 - Tracked supply-chain entries:
  `scripts/qa/offhost-shadow-restore-runner.template.sh`,
  `scripts/qa/validate-offhost-shadow-restore-attachments.py` and
@@ -1349,6 +1356,17 @@ validate_copied_batch(private/batch, source_metadata)
  environment values or stderr.
 - Run `dump_list` and target confirmation independently before the existing
  restore script. A failed stage leaves every later stage `NOT_RUN`.
+- Measure the eight data-restore stages in this exact continuous order:
+  `batch_copy` -> `postgres_container_start` -> `dump_list` ->
+  `target_confirmation` -> `restore_apply` -> `database_verify` ->
+  `storage_attachment_verify` -> `final_integrity`. One timestamp closes the
+  current stage and starts the next, so no elapsed time is hidden between
+  fields. The total is computed from the measured stage fields and must equal
+  their exact sum.
+- On failure, finalize the active stage and emit the completed/current numeric
+  durations plus later `NOT_RUN` fields before runtime teardown. A pre-copy
+  failure emits `NOT_RUN` for the total and every stage. Do not emit the total
+  a second time on the success-only summary path.
 - For the official PostgreSQL image, target readiness must use a real query over
  container-local TCP loopback (`psql -h 127.0.0.1`), not default Unix-socket
  `pg_isready`. The initdb temporary server listens only on its socket, while
@@ -1399,7 +1417,8 @@ validate_copied_batch(private/batch, source_metadata)
  `DUMP_LIST_NONZERO` / `dump_list_failed`.
 - Bounded TCP target queries remain non-zero -> `target_confirmation=FAIL`,
  apply `NOT_RUN`, category `TARGET_QUERY_NONZERO`; preserve the last numeric
-  exit, attempt count and zero success streak.
+  exit, attempt count and zero success streak. All query retries and their
+  one-second intervals accrue only to `rto_target_confirmation_seconds`.
 - Target query succeeds but the exact shadow name differs -> fail immediately
  with `target_name_match=no`, category `TARGET_NAME_MISMATCH`; do not retry a
  confirmed wrong target.
@@ -1411,6 +1430,10 @@ validate_copied_batch(private/batch, source_metadata)
  attachment checks still decide the final recovery result.
 - Diagnostics emitted only after cleanup, or cleanup residue non-zero -> the
  attempt is non-clean and cannot be reported as `GO`.
+- A measured stage is negative/non-numeric, a completed stage is `NOT_RUN`, a
+  later stage is numeric after an earlier failure, or the stage sum differs
+  from `data_restore_rto_seconds` -> runner contract failure; do not render or
+  upload the candidate.
 - Missing/extra render key, newline/shell syntax, unsafe path/name, malformed
  digest or pre-existing output -> render fails without publishing a partial
  candidate.
@@ -1434,6 +1457,12 @@ validate_copied_batch(private/batch, source_metadata)
  migration and attachment verification.
 - Good: Unix-socket availability alone does not pass; loopback TCP becomes
  available, two exact target queries succeed consecutively, then apply runs.
+- Good: an apply failure records numeric copy/container/list/target/apply
+  durations, leaves database/storage/final verification `NOT_RUN`, emits their
+  exact partial sum, and only then tears down the private runtime.
+- Base: a successful run emits all eight numeric durations once and their sum
+  exactly equals `data_restore_rto_seconds`; the value remains a data-only
+  exercise RTO, not a complete OA recovery RTO.
 - Bad: emit only `database_restore_failed` after deleting the sole private log.
 - Bad: parse and publish arbitrary pg_restore lines, or treat a non-zero exit as
  success because the restored database appears partly populated.
@@ -1450,6 +1479,15 @@ validate_copied_batch(private/batch, source_metadata)
  with an equivalent test implementation.
 - Inject list failure, target mismatch and list-pass/apply-failure. Assert exact
  stage enums, later-stage `NOT_RUN`, execution path and numeric exit code.
+- Drive the rendered runner's timing functions with deterministic timestamps.
+  Assert all eight successful stage durations and their exact total; inject
+  target-confirmation and restore-apply failures and assert completed/current
+  numeric values, later `NOT_RUN`, one total field and timing output before the
+  runtime cleanup marker.
+- Statically assert that the bounded TCP confirmation call is enclosed by the
+  `dump_list -> target_confirmation` and
+  `target_confirmation -> restore_apply` timing transitions; a retry sleep
+  must not be charged to container start or restore apply.
 - Extract the target-confirmation function from the rendered runner and drive
  it with: socket-only availability/TCP failure, first-query recovery, streak
  reset, exact-name mismatch and permanent bounded failure. Assert every probe
@@ -1485,6 +1523,7 @@ validate_copied_batch(private/batch, source_metadata)
 fixture: wait for two queries
 runner: first socket pg_isready -> continue
 restore-db > private.log || fail database_restore_failed
+RTO = finish - start; stage gaps are inferred later
 finally: delete private runtime
 print result=NO_GO
 ```
@@ -1494,8 +1533,9 @@ print result=NO_GO
 ```text
 tracked template + exact private config -> deterministic candidate + SHA-256
 contract drives rendered TCP target function -> list -> target -> real restore path
-capture only stage/path/exit/category enums
-emit fixed diagnostics -> finally teardown raw private runtime
+continuous stage transitions -> exact stage sum + one data-only RTO
+capture only stage/path/exit/category/duration allowlists
+finalize active stage -> emit fixed diagnostics -> finally teardown raw private runtime
 non-zero apply remains FAIL
 ```

@@ -130,14 +130,16 @@
 - [x] 最新批次来源与当前生产构建相同，探针时年龄 `64625` 秒且不超过 24 小时；run id 配对、DB/附件大小与 SHA-256、source digest 均通过。
 - [x] 来源/当前 migration 目录为 `65/65`、差异 `0`；仓库 HEAD 相对生产构建的应用代码漂移文件数 `0`。
 - [x] 探针 root 入口与普通用户暂存残留均为 `0`；按已验证私有配置取得的 dead-man/健康 timer 实际 unit 状态为 `active/active`。
- [ ] RTO 八阶段实现尚未批准，因此未渲染 Attempt 5 最终候选、未生成最终候选 SHA-256，也未执行实际恢复。
+- [x] 本次探针完成时 RTO 八阶段实现尚未批准，因此当时未渲染 Attempt 5 最终候选、未生成最终候选 SHA-256，也未执行实际恢复；后续本地实现授权单独追加，不改写该时点事实。

-### Attempt 5 前 RTO 分阶段提案（待单独批准实现）
+### Attempt 5 前 RTO 分阶段实施（已批准仅本地实现）

 - [x] 设计口径已固定为总 RTO 加八段串行秒数：批次复制、容器启动、dump list、目标确认、restore apply、migration/12 类表、附件恢复与关联校验、最终源/副本复核。
- [ ] runner 增加固定 allowlist 阶段字段，成功时八段均为非负整数且精确加总为 `data_restore_rto_seconds`；失败时已完成段和当前失败段保留数字，后续段为 `NOT_RUN`。
- [ ] 合同注入各阶段成功/失败，验证最长 TCP 轮询只进入目标确认阶段、后续 `NOT_RUN`、总数精确加总、输出不含路径/SQL/凭据；原有清理与恢复合同继续通过。
- [ ] 未经单独批准，不修改 runner、重新渲染 Attempt 5 候选或执行实际恢复。
+- [x] runner 增加固定 allowlist 阶段字段，成功时八段均为非负整数且精确加总为 `data_restore_rto_seconds`；失败时已完成段和当前失败段保留数字，后续段为 `NOT_RUN`。
+- [x] 合同注入各阶段成功/失败，验证最长 TCP 轮询只进入目标确认阶段、后续 `NOT_RUN`、总数精确加总、输出不含路径/SQL/凭据；原有清理与恢复合同继续通过。
+- [x] 用户批准的当前边界仅为本地修改 runner、更新可复跑合同并执行回归；不连接 ARM、不重新渲染 Attempt 5 现场候选、不执行实际恢复。
+- [ ] 本轮八阶段计时与完整合同通过后，下一份满足 24 小时门禁的新鲜批次直接进入 Attempt 5 的候选渲染、SHA-256 核对、只读探针和实际恢复审批流程。此后发现的改进项，除非会导致 Attempt 5 失败或产生错误结论，否则只登记为后续任务，不再作为 Attempt 5 前置增强插队。
+- [x] 八阶段状态必须在失败路径清理前输出；阶段边界沿用 `dump_list`、`target_confirmation`、`restore_apply` 等既有诊断词汇，不建立第二套同义命名。

 ## 验证命令类别

@@ -452,3 +452,17 @@ Attempt 4 前后探针的最新构建短标识、run id 配对、产物完整性
 timer 状态只采用已验证私有配置中的实际 unit 名只读取得；一次使用猜测名称得到的无效读数已当场撤销，未进入证据结论。探针结果表明该最新批次继续满足构建、新鲜度、身份和完整性前置门禁，但仍不是恢复成功证明。

 Attempt 5 最终候选尚未渲染。原因是用户提出的 RTO 分阶段记录已进入设计、但 runner 实现尚未另行批准；若先渲染，后续计时代码会改变候选字节和 SHA-256。必须先裁决并完成 RTO 八阶段本地合同，再从跟踪模板重新确定性渲染候选并核对 SHA-256，实际恢复仍须单独批准。
+
+## Attempt 5 前八阶段 RTO 本地合同
+
+日期：2026-08-25
+
+用户随后批准的范围仅为本地实现八阶段 RTO 计时并执行合同回归；本阶段未连接 ARM、未渲染现场候选、未生成 Attempt 5 候选 SHA-256，也未执行实际恢复。此前探针批次的 `64625` 秒新鲜度只属于探针时点，后续不得复用，实际恢复前必须重新选择最新正式批次并重跑只读身份门禁。
+
+runner 现在以连续边界依次记录 `batch_copy`、`postgres_container_start`、`dump_list`、`target_confirmation`、`restore_apply`、`database_verify`、`storage_attachment_verify` 和 `final_integrity`。同一个时间点关闭前一段并开启后一段；成功路径八段均为非负整数且精确加总为唯一的 `data_restore_rto_seconds`。失败路径先结算当前阶段，保留已完成和当前阶段的数字，后续阶段为 `NOT_RUN`，然后才清理私有 runtime。最长 60 次 TCP 目标确认及其 1 秒间隔只位于 `target_confirmation` 边界内。
+
+可复跑合同直接驱动 rendered runner 的同一计时函数。确定性成功夹具得到八段 `3/5/2/11/7/13/17/3` 秒，总计 `61` 秒且总字段只输出一次；目标确认失败夹具保留 `2/4/1/9` 秒、后四段 `NOT_RUN`、部分总计 `16` 秒，并证明计时输出位于 runtime cleanup 之前；restore apply 失败夹具记录当前段 `12` 秒、后续三段 `NOT_RUN`、部分总计 `28` 秒。捕获输出不含路径、连接串、凭据或 SQL。
+
+完整本地门禁结果：runner 合同全部 `PASS`；真实 PostgreSQL 16 fixture 和未修改的 `scripts/restore-db.sh` Docker stdin/host-file 双路径继续通过；既有成对备份/恢复演练 `PASS`；自动异机备份合同 `74/74 PASS`；task validate、Trellis governance、process baseline、docs registry、敏感扫描和 `git diff --check` 通过。文档审计错误 `0`，保留既有未登记文件 warning；本地测试容器残留 `0`，`scripts/restore-db.sh` 相对生产构建仍无字节漂移。
+
+Attempt 5 前增强停止线从本节起生效：八阶段计时完成并通过合同后，下一份满足 24 小时门禁的新鲜批次直接进入最终候选渲染、SHA-256 核对、只读探针和实际恢复审批。此后发现的改进项，除非会导致 Attempt 5 失败或产生错误结论，否则只登记为后续任务，不再作为 Attempt 5 前置增强插队。本地合同通过仍只证明演练器合同，不回答最新正式备份是否可恢复。
@@ -29,7 +29,18 @@ CONTAINER_NAME=""
 RUNTIME_CLEANED=0
 RESULT_STATUS="NO_GO"
 RPO_SECONDS=""
-RTO_SECONDS=""
+RTO_SECONDS="NOT_RUN"
+RTO_BATCH_COPY_SECONDS="NOT_RUN"
+RTO_POSTGRES_CONTAINER_START_SECONDS="NOT_RUN"
+RTO_DUMP_LIST_SECONDS="NOT_RUN"
+RTO_TARGET_CONFIRMATION_SECONDS="NOT_RUN"
+RTO_RESTORE_APPLY_SECONDS="NOT_RUN"
+RTO_DATABASE_VERIFY_SECONDS="NOT_RUN"
+RTO_STORAGE_ATTACHMENT_VERIFY_SECONDS="NOT_RUN"
+RTO_FINAL_INTEGRITY_SECONDS="NOT_RUN"
+RTO_ACTIVE_STAGE=""
+RTO_STAGE_STARTED_EPOCH=""
+RTO_FINALIZED=0
 MIGRATION_SUCCESS=""
 MIGRATION_OTHER=""
 MIGRATION_SET_MATCH=""
@@ -66,6 +77,94 @@ require_command() {
  command -v "$1" >/dev/null 2>&1 || fail "required_command_missing"
 }

+rto_now() {
+  date -u +%s
+}
+
+rto_set_stage_seconds() {
+  local stage="$1"
+  local seconds="$2"
+  case "$stage" in
+    batch_copy) RTO_BATCH_COPY_SECONDS="$seconds" ;;
+    postgres_container_start) RTO_POSTGRES_CONTAINER_START_SECONDS="$seconds" ;;
+    dump_list) RTO_DUMP_LIST_SECONDS="$seconds" ;;
+    target_confirmation) RTO_TARGET_CONFIRMATION_SECONDS="$seconds" ;;
+    restore_apply) RTO_RESTORE_APPLY_SECONDS="$seconds" ;;
+    database_verify) RTO_DATABASE_VERIFY_SECONDS="$seconds" ;;
+    storage_attachment_verify) RTO_STORAGE_ATTACHMENT_VERIFY_SECONDS="$seconds" ;;
+    final_integrity) RTO_FINAL_INTEGRITY_SECONDS="$seconds" ;;
+    *) return 1 ;;
+  esac
+}
+
+rto_begin() {
+  local stage="$1"
+  local now
+  [[ -z "$RTO_ACTIVE_STAGE" && "$RTO_FINALIZED" -eq 0 ]] || return 1
+  now="$(rto_now)"
+  [[ "$now" =~ ^[0-9]+$ ]] || return 1
+  rto_set_stage_seconds "$stage" "NOT_RUN" || return 1
+  RTO_ACTIVE_STAGE="$stage"
+  RTO_STAGE_STARTED_EPOCH="$now"
+}
+
+rto_close_active_at() {
+  local now="$1"
+  local elapsed
+  [[ -n "$RTO_ACTIVE_STAGE" && "$RTO_STAGE_STARTED_EPOCH" =~ ^[0-9]+$ \
+    && "$now" =~ ^[0-9]+$ && "$now" -ge "$RTO_STAGE_STARTED_EPOCH" ]] \
+    || return 1
+  elapsed="$((now - RTO_STAGE_STARTED_EPOCH))"
+  rto_set_stage_seconds "$RTO_ACTIVE_STAGE" "$elapsed" || return 1
+}
+
+rto_transition() {
+  local expected_stage="$1"
+  local next_stage="$2"
+  local now
+  [[ "$RTO_ACTIVE_STAGE" == "$expected_stage" && "$RTO_FINALIZED" -eq 0 ]] \
+    || return 1
+  now="$(rto_now)"
+  rto_close_active_at "$now" || return 1
+  rto_set_stage_seconds "$next_stage" "NOT_RUN" || return 1
+  RTO_ACTIVE_STAGE="$next_stage"
+  RTO_STAGE_STARTED_EPOCH="$now"
+}
+
+rto_finalize() {
+  local now
+  local value
+  local total=0
+  local measured=0
+  if [[ "$RTO_FINALIZED" -eq 1 ]]; then
+    return 0
+  fi
+  if [[ -n "$RTO_ACTIVE_STAGE" ]]; then
+    now="$(rto_now)"
+    rto_close_active_at "$now" || return 1
+    RTO_ACTIVE_STAGE=""
+    RTO_STAGE_STARTED_EPOCH=""
+  fi
+  for value in \
+    "$RTO_BATCH_COPY_SECONDS" \
+    "$RTO_POSTGRES_CONTAINER_START_SECONDS" \
+    "$RTO_DUMP_LIST_SECONDS" \
+    "$RTO_TARGET_CONFIRMATION_SECONDS" \
+    "$RTO_RESTORE_APPLY_SECONDS" \
+    "$RTO_DATABASE_VERIFY_SECONDS" \
+    "$RTO_STORAGE_ATTACHMENT_VERIFY_SECONDS" \
+    "$RTO_FINAL_INTEGRITY_SECONDS"; do
+    if [[ "$value" =~ ^[0-9]+$ ]]; then
+      total="$((total + value))"
+      measured=1
+    fi
+  done
+  if [[ "$measured" -eq 1 ]]; then
+    RTO_SECONDS="$total"
+  fi
+  RTO_FINALIZED=1
+}
+
 confirm_postgres_target() {
  local actual_database=""
  local query_exit_code=0
@@ -182,6 +281,15 @@ emit_restore_diagnostics() {
  echo "pg_restore_exit_code=$PG_RESTORE_EXIT_CODE"
  echo "pg_restore_error_count=$PG_RESTORE_ERROR_COUNT"
  echo "restore_error_category=$RESTORE_ERROR_CATEGORY"
+  echo "rto_batch_copy_seconds=$RTO_BATCH_COPY_SECONDS"
+  echo "rto_postgres_container_start_seconds=$RTO_POSTGRES_CONTAINER_START_SECONDS"
+  echo "rto_dump_list_seconds=$RTO_DUMP_LIST_SECONDS"
+  echo "rto_target_confirmation_seconds=$RTO_TARGET_CONFIRMATION_SECONDS"
+  echo "rto_restore_apply_seconds=$RTO_RESTORE_APPLY_SECONDS"
+  echo "rto_database_verify_seconds=$RTO_DATABASE_VERIFY_SECONDS"
+  echo "rto_storage_attachment_verify_seconds=$RTO_STORAGE_ATTACHMENT_VERIFY_SECONDS"
+  echo "rto_final_integrity_seconds=$RTO_FINAL_INTEGRITY_SECONDS"
+  echo "data_restore_rto_seconds=$RTO_SECONDS"
 }

 finish() {
@@ -191,6 +299,10 @@ finish() {
  trap - EXIT HUP INT TERM
  # Emit only allowlisted stage diagnostics while the private runtime still
  # exists. Raw stderr remains mode 0600 and is removed by cleanup_runtime.
+  if ! rto_finalize; then
+    FAILURE_CATEGORY="rto_stage_finalize_failed"
+    exit_code=1
+  fi
  emit_restore_diagnostics
  if [[ "$RUNTIME_CLEANED" -ne 1 ]]; then
    cleanup_runtime || runtime_cleanup="FAIL"
@@ -204,7 +316,6 @@ finish() {
  if [[ "$exit_code" -eq 0 ]]; then
    echo "result=$RESULT_STATUS"
    echo "rpo_seconds=$RPO_SECONDS"
-    echo "data_restore_rto_seconds=$RTO_SECONDS"
    echo "migration_success=$MIGRATION_SUCCESS"
    echo "migration_other=$MIGRATION_OTHER"
    echo "migration_set_match=$MIGRATION_SET_MATCH"
@@ -412,11 +523,13 @@ mkdir -m 0700 "$WORK_DIR"
 private_log="$WORK_DIR/private.log"
 : >"$private_log"
 chmod 0600 "$private_log"
-rto_started_epoch="$(date -u +%s)"
+rto_begin batch_copy || fail "rto_stage_begin_failed"
 cp -a -- "$selected_batch" "$WORK_DIR/batch"
 copied_metadata="$(validate_batch "$WORK_DIR/batch" copy "$SOURCE_RUN_ID")" \
  || fail "copied_batch_invalid"
 [[ "$copied_metadata" == "$batch_metadata" ]] || fail "copied_batch_identity_mismatch"
+rto_transition batch_copy postgres_container_start \
+  || fail "rto_stage_transition_failed"

 env_file="$WORK_DIR/postgres.env"
 postgres_password="$(openssl rand -hex 24)"
@@ -429,6 +542,8 @@ chmod 0600 "$env_file"
 printf '%s\n' '[stage=postgres_container_start]' >>"$private_log"
 docker run -d --rm --name "$CONTAINER_NAME" --network none --env-file "$env_file" \
  "$POSTGRES_IMAGE" >>"$private_log" 2>&1 || fail "postgres_container_start_failed"
+rto_transition postgres_container_start dump_list \
+  || fail "rto_stage_transition_failed"

 PHASE="restore_database"
 DUMP_LIST_STATUS="FAIL"
@@ -444,6 +559,8 @@ if [[ "$DUMP_LIST_EXIT_CODE" -ne 0 ]]; then
  fail "dump_list_failed"
 fi
 DUMP_LIST_STATUS="PASS"
+rto_transition dump_list target_confirmation \
+  || fail "rto_stage_transition_failed"

 TARGET_CONFIRMATION_STATUS="FAIL"
 printf '%s\n' '[stage=target_confirmation]' >>"$private_log"
@@ -452,6 +569,8 @@ if ! confirm_postgres_target; then
  fail "restore_target_confirmation_failed"
 fi
 TARGET_CONFIRMATION_STATUS="PASS"
+rto_transition target_confirmation restore_apply \
+  || fail "rto_stage_transition_failed"

 real_docker_path="$(command -v docker)"
 docker_proxy_dir="$WORK_DIR/docker-proxy"
@@ -530,6 +649,8 @@ if [[ "$RESTORE_SCRIPT_EXIT_CODE" -ne 0 || "$PG_RESTORE_EXIT_CODE" -ne 0 ]]; the
 fi
 RESTORE_APPLY_STATUS="PASS"
 RESTORE_ERROR_CATEGORY="NONE"
+rto_transition restore_apply database_verify \
+  || fail "rto_stage_transition_failed"

 PHASE="verify_database"
 migration_names="$(PGPASSWORD="$postgres_password" docker exec -i -e PGPASSWORD \
@@ -581,6 +702,8 @@ SQL
 )" || fail "table_count_query_failed"
 [[ "$(printf '%s\n' "$TABLE_COUNTS" | sed '/^$/d' | wc -l | tr -d ' ')" == "12" ]] \
  || fail "table_count_key_mismatch"
+rto_transition database_verify storage_attachment_verify \
+  || fail "rto_stage_transition_failed"

 PHASE="restore_storage"
 restored_storage="$WORK_DIR/restored-storage"
@@ -613,6 +736,8 @@ SQL
 read -r ATTACHMENT_TOTAL ATTACHMENT_MISMATCH attachment_s3_count <<<"$attachment_result"
 [[ "$ATTACHMENT_TOTAL" =~ ^[0-9]+$ && "$ATTACHMENT_MISMATCH" == "0" \
  && "$attachment_s3_count" == "0" ]] || fail "attachment_reference_mismatch"
+rto_transition storage_attachment_verify final_integrity \
+  || fail "rto_stage_transition_failed"

 PHASE="post_validation"
 copied_metadata_after="$(validate_batch "$WORK_DIR/batch" copy "$SOURCE_RUN_ID")" \
@@ -621,8 +746,7 @@ source_metadata_after="$(validate_batch "$selected_batch" source "")" \
  || fail "source_archive_postcheck_failed"
 [[ "$copied_metadata_after" == "$batch_metadata" \
  && "$source_metadata_after" == "$batch_metadata" ]] || fail "archive_identity_changed"
-rto_finished_epoch="$(date -u +%s)"
-RTO_SECONDS="$((rto_finished_epoch - rto_started_epoch))"
+rto_finalize || fail "rto_stage_finalize_failed"

 deadman_enabled_after="$(systemctl is-enabled "$DEADMAN_TIMER_UNIT" 2>/dev/null || true)"
 deadman_active_after="$(systemctl is-active "$DEADMAN_TIMER_UNIT" 2>/dev/null || true)"
@@ -432,14 +432,17 @@ finish = re.search(r"\nfinish\(\) \{(?P<body>.*?)\n\}\ntrap finish", text, re.S)
 if not finish:
    raise SystemExit("finish function not found")
 body = finish.group("body")
+finalize_position = body.find("\n  if ! rto_finalize; then")
 emit_position = body.find("\n  emit_restore_diagnostics")
 cleanup_position = body.find("\n    cleanup_runtime ||")
+if finalize_position == -1:
+    raise SystemExit("RTO finalization missing")
 if emit_position == -1:
    raise SystemExit("diagnostic emission missing")
 if cleanup_position == -1:
    raise SystemExit("runtime cleanup call missing")
-if emit_position > cleanup_position:
-    raise SystemExit("diagnostics are emitted after cleanup")
+if not (finalize_position < emit_position < cleanup_position):
+    raise SystemExit("RTO/stage diagnostics are not finalized before cleanup")
 required = {
    'echo "dump_list=$DUMP_LIST_STATUS"',
    'echo "target_confirmation=$TARGET_CONFIRMATION_STATUS"',
@@ -452,14 +455,41 @@ required = {
    'echo "pg_restore_exit_code=$PG_RESTORE_EXIT_CODE"',
    'echo "pg_restore_error_count=$PG_RESTORE_ERROR_COUNT"',
    'echo "restore_error_category=$RESTORE_ERROR_CATEGORY"',
+    'echo "rto_batch_copy_seconds=$RTO_BATCH_COPY_SECONDS"',
+    'echo "rto_postgres_container_start_seconds=$RTO_POSTGRES_CONTAINER_START_SECONDS"',
+    'echo "rto_dump_list_seconds=$RTO_DUMP_LIST_SECONDS"',
+    'echo "rto_target_confirmation_seconds=$RTO_TARGET_CONFIRMATION_SECONDS"',
+    'echo "rto_restore_apply_seconds=$RTO_RESTORE_APPLY_SECONDS"',
+    'echo "rto_database_verify_seconds=$RTO_DATABASE_VERIFY_SECONDS"',
+    'echo "rto_storage_attachment_verify_seconds=$RTO_STORAGE_ATTACHMENT_VERIFY_SECONDS"',
+    'echo "rto_final_integrity_seconds=$RTO_FINAL_INTEGRITY_SECONDS"',
+    'echo "data_restore_rto_seconds=$RTO_SECONDS"',
 }
 missing = sorted(value for value in required if value not in text)
 if missing:
    raise SystemExit(f"diagnostic fields missing: {missing}")
 if 'RESTORE_ERROR_CATEGORY="RESTORE_APPLY_NONZERO"' not in text:
    raise SystemExit("non-zero apply classification missing")
+transitions = [
+    "rto_begin batch_copy",
+    "rto_transition batch_copy postgres_container_start",
+    "rto_transition postgres_container_start dump_list",
+    "rto_transition dump_list target_confirmation",
+    "rto_transition target_confirmation restore_apply",
+    "rto_transition restore_apply database_verify",
+    "rto_transition database_verify storage_attachment_verify",
+    "rto_transition storage_attachment_verify final_integrity",
+]
+positions = [text.find(value) for value in transitions]
+if any(position == -1 for position in positions) or positions != sorted(positions):
+    raise SystemExit("RTO stage transitions missing or out of order")
+target_call_position = text.find("if ! confirm_postgres_target; then")
+if not positions[3] < target_call_position < positions[4]:
+    raise SystemExit("TCP target confirmation is outside its RTO stage boundary")
+if text.count('echo "data_restore_rto_seconds=$RTO_SECONDS"') != 1:
+    raise SystemExit("data restore RTO must be emitted exactly once")
 PY
-pass "diagnostics precede finally-equivalent cleanup and use allowlisted fields"
+pass "eight-stage RTO and restore diagnostics precede cleanup with one vocabulary"

 cat >"$TEST_DIR/mock-target-docker" <<'MOCK_DOCKER'
 #!/usr/bin/env bash
@@ -738,6 +768,107 @@ CONTAINER_NAME=""
 sed -n '1,/^trap finish EXIT HUP INT TERM$/p' "$RUNNER" \
  | sed '$d' >"$TEST_DIR/runner-finalization-functions.sh"

+(
+  # shellcheck disable=SC1090
+  source "$TEST_DIR/runner-finalization-functions.sh"
+  RTO_TEST_NOW=100
+  rto_now() {
+    printf '%s\n' "$RTO_TEST_NOW"
+  }
+  rto_begin batch_copy
+  RTO_TEST_NOW=103
+  rto_transition batch_copy postgres_container_start
+  RTO_TEST_NOW=108
+  rto_transition postgres_container_start dump_list
+  RTO_TEST_NOW=110
+  rto_transition dump_list target_confirmation
+  RTO_TEST_NOW=121
+  rto_transition target_confirmation restore_apply
+  RTO_TEST_NOW=128
+  rto_transition restore_apply database_verify
+  RTO_TEST_NOW=141
+  rto_transition database_verify storage_attachment_verify
+  RTO_TEST_NOW=158
+  rto_transition storage_attachment_verify final_integrity
+  RTO_TEST_NOW=161
+  rto_finalize
+  emit_restore_diagnostics
+) >"$TEST_DIR/captured-rto-success" 2>&1
+for expected in \
+  'rto_batch_copy_seconds=3' \
+  'rto_postgres_container_start_seconds=5' \
+  'rto_dump_list_seconds=2' \
+  'rto_target_confirmation_seconds=11' \
+  'rto_restore_apply_seconds=7' \
+  'rto_database_verify_seconds=13' \
+  'rto_storage_attachment_verify_seconds=17' \
+  'rto_final_integrity_seconds=3' \
+  'data_restore_rto_seconds=61'; do
+  grep -Fxq "$expected" "$TEST_DIR/captured-rto-success" \
+    || fail "successful eight-stage RTO total or boundary mismatch"
+done
+[[ "$(grep -c '^data_restore_rto_seconds=' "$TEST_DIR/captured-rto-success")" -eq 1 ]] \
+  || fail "successful RTO total was not emitted exactly once"
+pass "rendered runner measures eight contiguous stages with an exact total"
+
+set +e
+(
+  # shellcheck disable=SC1090
+  source "$TEST_DIR/runner-finalization-functions.sh"
+  cleanup_runtime() {
+    echo "cleanup_marker=runtime"
+    return 0
+  }
+  cleanup_access() {
+    echo "cleanup_marker=access"
+    return 0
+  }
+  RTO_BATCH_COPY_SECONDS="2"
+  RTO_POSTGRES_CONTAINER_START_SECONDS="4"
+  RTO_DUMP_LIST_SECONDS="1"
+  RTO_ACTIVE_STAGE="target_confirmation"
+  RTO_STAGE_STARTED_EPOCH="200"
+  RTO_TEST_NOW=209
+  rto_now() {
+    printf '%s\n' "$RTO_TEST_NOW"
+  }
+  FAILURE_CATEGORY="restore_target_confirmation_failed"
+  DUMP_LIST_STATUS="PASS"
+  DUMP_LIST_EXIT_CODE="0"
+  TARGET_CONFIRMATION_STATUS="FAIL"
+  TARGET_CONFIRMATION_EXIT_CODE="1"
+  TARGET_CONFIRMATION_ATTEMPT_COUNT="60"
+  RESTORE_EXECUTION_PATH="DOCKER_STDIN"
+  RESTORE_ERROR_CATEGORY="TARGET_QUERY_NONZERO"
+  trap finish EXIT
+  exit 23
+) >"$TEST_DIR/captured-rto-target-failure" 2>&1
+rto_target_failure_exit="$?"
+set -e
+[[ "$rto_target_failure_exit" -eq 23 ]] \
+  || fail "target-confirmation timing failure did not preserve exit code"
+for expected in \
+  'rto_batch_copy_seconds=2' \
+  'rto_postgres_container_start_seconds=4' \
+  'rto_dump_list_seconds=1' \
+  'rto_target_confirmation_seconds=9' \
+  'rto_restore_apply_seconds=NOT_RUN' \
+  'rto_database_verify_seconds=NOT_RUN' \
+  'rto_storage_attachment_verify_seconds=NOT_RUN' \
+  'rto_final_integrity_seconds=NOT_RUN' \
+  'data_restore_rto_seconds=16'; do
+  grep -Fxq "$expected" "$TEST_DIR/captured-rto-target-failure" \
+    || fail "target-confirmation failure RTO boundary mismatch"
+done
+rto_line="$(grep -n '^rto_target_confirmation_seconds=9$' \
+  "$TEST_DIR/captured-rto-target-failure" | cut -d: -f1)"
+rto_cleanup_line="$(grep -n '^cleanup_marker=runtime$' \
+  "$TEST_DIR/captured-rto-target-failure" | cut -d: -f1)"
+[[ "$rto_line" =~ ^[0-9]+$ && "$rto_cleanup_line" =~ ^[0-9]+$ \
+  && "$rto_line" -lt "$rto_cleanup_line" ]] \
+  || fail "failed-stage RTO was not emitted before teardown"
+pass "target-confirmation failure preserves elapsed stages and later NOT_RUN before cleanup"
+
 capture_finalization_case() {
  local output_file="$1"
  local failure_category="$2"
@@ -805,7 +936,13 @@ grep -Fxq 'target_confirmation=NOT_RUN' \
  || fail "target stage ran after dump-list failure"
 grep -Fxq 'restore_apply=NOT_RUN' "$TEST_DIR/captured-dump-list-failure" \
  || fail "apply stage ran after dump-list failure"
-pass "dump-list failure keeps target and apply stages NOT_RUN"
+grep -Fxq 'rto_batch_copy_seconds=NOT_RUN' \
+  "$TEST_DIR/captured-dump-list-failure" \
+  || fail "pre-RTO failure unexpectedly reported a measured stage"
+grep -Fxq 'data_restore_rto_seconds=NOT_RUN' \
+  "$TEST_DIR/captured-dump-list-failure" \
+  || fail "pre-RTO failure unexpectedly reported a measured total"
+pass "pre-RTO failure keeps timing, target and apply fields NOT_RUN"

 capture_finalization_case \
  "$TEST_DIR/captured-target-failure" \
@@ -849,6 +986,16 @@ set +e
  PG_RESTORE_EXIT_CODE="23"
  PG_RESTORE_ERROR_COUNT="UNKNOWN"
  RESTORE_ERROR_CATEGORY="RESTORE_APPLY_NONZERO"
+  RTO_BATCH_COPY_SECONDS="2"
+  RTO_POSTGRES_CONTAINER_START_SECONDS="4"
+  RTO_DUMP_LIST_SECONDS="1"
+  RTO_TARGET_CONFIRMATION_SECONDS="9"
+  RTO_ACTIVE_STAGE="restore_apply"
+  RTO_STAGE_STARTED_EPOCH="300"
+  RTO_TEST_NOW=312
+  rto_now() {
+    printf '%s\n' "$RTO_TEST_NOW"
+  }
  trap finish EXIT
  exit 23
 ) >"$TEST_DIR/captured-finalization" 2>&1
@@ -869,7 +1016,13 @@ grep -Fxq 'temporary_access_cleanup=PASS' "$TEST_DIR/captured-finalization" \
  || fail "access teardown result was not retained"
 grep -Fxq 'pg_restore_exit_code=23' "$TEST_DIR/captured-finalization" \
  || fail "pg_restore exit code disappeared after teardown"
-pass "injected failure emits readable stages before finally-equivalent teardown"
+grep -Fxq 'rto_restore_apply_seconds=12' "$TEST_DIR/captured-finalization" \
+  || fail "restore-apply failure duration missing"
+grep -Fxq 'rto_database_verify_seconds=NOT_RUN' "$TEST_DIR/captured-finalization" \
+  || fail "database verification ran after restore-apply failure"
+grep -Fxq 'data_restore_rto_seconds=28' "$TEST_DIR/captured-finalization" \
+  || fail "restore-apply failure RTO total mismatch"
+pass "injected restore-apply failure preserves diagnostics and RTO before teardown"

 if grep -Eiq 'postgresql://|SELECT[[:space:]]|/tmp/|local_fixture_password' \
  "$TEST_DIR/captured-dump-list-failure" \