chore(scripts): add manual transition follow-up monitor (#5328)

* test(scripts): cover manual transition runbooks

Add a script-test entry for the manual transition rollout and soak runbook generators. The test covers ratio fail-fast behavior, generated status/metrics/journal checks, and runnable dry-run artifacts.

Verification:

- ./scripts/test_manual_transition_runbooks.sh

- make script-tests

Co-Authored-By: heihutu <heihutu@gmail.com>

* test(scripts): assert nightly stress snapshot hooks

Co-Authored-By: heihutu <heihutu@gmail.com>

* feat(scripts): add transition CI follow-up monitor

Co-Authored-By: heihutu <heihutu@gmail.com>

---------

Co-authored-by: heihutu <heihutu@gmail.com>
This commit is contained in:
houseme
2026-07-27 16:36:41 +08:00
committed by GitHub
parent bc2d8e0c60
commit 464bf45e15
4 changed files with 391 additions and 7 deletions
+1
View File
@@ -90,6 +90,7 @@ their issue closes.
| `manual_transition_journal_audit.sh` | dev-tool | Journal + metrics + log audit for manual transition jobs | — |
| `manual_transition_mixed_rollout_matrix.sh` | dev-tool | Matrix generator for mixed-version rollout phases | — |
| `manual_transition_mixed_rollout_runbook.sh` | dev-tool | Reusable mixed-version rollout runbook generator (external run) | — |
| `monitor_manual_transition_ci.sh` | dev-tool | CI workflow/status watcher for manual transition follow-up monitoring | — |
| `manual_transition_soak_matrix.sh` | dev-tool | Matrix generator for nightly stress windows | — |
| `manual_transition_nightly_stress_runbook.sh` | dev-tool | Nightly stress entrypoint with failure snapshot templates | — |
| `install-flatc.sh` | dev-tool | Local flatc installer (macOS) | — |
@@ -297,6 +297,8 @@ run_entry() {
local response
local job_id
local status
local status_json
local status_info
local headers=()
if [[ "$budget_status" == "over-budget" ]]; then
@@ -304,10 +306,6 @@ run_entry() {
return 0
fi
if [[ "$UNKNOWN_FAILURE_COUNT_THRESHOLD" -gt 0 && "$expected_ops" -gt "$UNKNOWN_FAILURE_COUNT_THRESHOLD" ]]; then
:
fi
prefix="${JOB_PREFIX}/${tag}/${size}/${read_pct}r${write_pct}w"
query="bucket=$(url_encode "$JOB_BUCKET")"
query="${query}&prefix=$(url_encode "$prefix")"
@@ -335,10 +333,77 @@ run_entry() {
return 1
fi
status="$(curl -sS "${headers[@]}" -X GET "${ENDPOINT%/}/rustfs/admin/v3/ilm/transition/jobs/${job_id}" | jq -r '.failure_reason // empty')"
if [[ -n "$status" && "$status" != "null" ]]; then
snapshot_failure "$tag" "failure_reason=${status}" "$job_id"
if ! status_json="$(curl -sS "${headers[@]}" -X GET "${ENDPOINT%/}/rustfs/admin/v3/ilm/transition/jobs/${job_id}")"; then
snapshot_failure "$tag" "job_status_fetch_failed" "$job_id"
return 1
fi
if ! status_info="$(printf '%s' "$status_json" | jq -r '[.status // "", .failure_reason // "", (.report.enqueued // 0), (.report.transition_completed // 0), (.report.transition_failed // 0), (.report.tier_failure_by_reason["unknown"] // 0), (.queue_snapshot.queued // 0), (.queue_snapshot.active // 0), (.queue_snapshot.compensation_pending // 0), (.queue_snapshot.compensation_running // 0), (.queue_snapshot.queue_full // 0), (.queue_snapshot.queue_send_timeout // 0)] | map(tostring) | join(\"\\u0000\")')"; then
snapshot_failure "$tag" "invalid_job_status_json" "$job_id"
return 1
fi
IFS=$'\0' read -r status failure_reason report_enqueued report_completed report_failed report_unknown_failure queue_snapshot_queued queue_snapshot_active compensation_pending compensation_running queue_full queue_send_timeout <<< "$status_info"
if [[ -n "$failure_reason" && "$failure_reason" != "null" ]]; then
snapshot_failure "$tag" "failure_reason=${failure_reason}" "$job_id"
fi
# Threshold checks are applied to terminal job statuses only to avoid transient noise
# from active windows that are still processing.
if is_terminal_status "$status"; then
local mismatch_count
local mismatch_ratio
local unknown_ratio
local ratio_denominator
mismatch_count="$((report_enqueued - report_completed - report_failed))"
if (( mismatch_count < 0 )); then
mismatch_count=0
fi
mismatch_count="$((queue_snapshot_queued + queue_snapshot_active + compensation_pending + compensation_running + queue_full + queue_send_timeout + mismatch_count))"
if (( expected_ops > 0 )); then
ratio_denominator="$expected_ops"
else
ratio_denominator="$report_enqueued"
fi
if (( ratio_denominator <= 0 )); then
ratio_denominator=0
fi
if (( ratio_denominator > 0 )); then
unknown_ratio="$(awk -v unknown="$report_unknown_failure" -v denom="$ratio_denominator" 'BEGIN{printf "%.6f", (unknown / denom)}')"
mismatch_ratio="$(awk -v mismatch="$mismatch_count" -v denom="$ratio_denominator" 'BEGIN{printf "%.6f", (mismatch / denom)}')"
if (( UNKNOWN_FAILURE_RATIO_THRESHOLD > 0 )) && \
awk "BEGIN{exit !($unknown_ratio > ${UNKNOWN_FAILURE_RATIO_THRESHOLD})}"; then
snapshot_failure "$tag" "unknown_failure_ratio=${unknown_ratio}/expected=${ratio_denominator}/threshold=${UNKNOWN_FAILURE_RATIO_THRESHOLD}" "$job_id"
fi
if (( UNKNOWN_FAILURE_COUNT_THRESHOLD > 0 )) && (( report_unknown_failure > UNKNOWN_FAILURE_COUNT_THRESHOLD )); then
snapshot_failure "$tag" "unknown_failure_count=${report_unknown_failure}/threshold=${UNKNOWN_FAILURE_COUNT_THRESHOLD}" "$job_id"
fi
if (( QUEUE_MISMATCH_RATIO_THRESHOLD > 0 )) && \
awk "BEGIN{exit !($mismatch_ratio > ${QUEUE_MISMATCH_RATIO_THRESHOLD})}"; then
snapshot_failure "$tag" "queue_mismatch_ratio=${mismatch_ratio}/expected=${ratio_denominator}/threshold=${QUEUE_MISMATCH_RATIO_THRESHOLD}" "$job_id"
fi
fi
fi
# Keep compatibility with older callers that monitor queue drift via full terminal status.
curl -sS "${headers[@]}" -X GET "${ENDPOINT%/}/rustfs/admin/v3/ilm/transition/jobs/${job_id}" | jq '{status, report, queue_snapshot, failure_reason}'
}
is_terminal_status() {
local current_status="$1"
case "$current_status" in
completed|partial|failed|cancelled|unknown)
return 0
;;
*)
return 1
;;
esac
}
main() {
+294
View File
@@ -0,0 +1,294 @@
#!/usr/bin/env bash
# Copyright 2026 RustFS Team
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
set -euo pipefail
usage() {
cat <<'USAGE'
Usage:
scripts/monitor_manual_transition_ci.sh [options]
Optional:
--repo Target repo for gh run listing (default: rustfs/rustfs)
--workflows Comma-separated workflow file names (default: ci.yml,e2e-s3tests.yml,e2e-replication-nightly.yml,mint.yml,fuzz.yml)
--branch Branch to check (default: main)
--runs Number of latest runs to sample per workflow (default: 5)
--require-complete Fail if any sampled run is still queued/in_progress/running
--pr Check PR checks for a specific PR number in --repo
--issues Comma-separated issue numbers for markdown follow-up snapshot
--help
This helper is read-only and does not mutate repository state.
USAGE
}
require_cmd() {
local cmd="$1"
if ! command -v "$cmd" >/dev/null 2>&1; then
echo "ERROR: required command not found: $cmd" >&2
exit 1
fi
}
arg_value() {
local flag="$1"
local value="${2:-}"
if [[ -z "$value" || "$value" == --* ]]; then
echo "ERROR: missing value for $flag" >&2
exit 1
fi
printf '%s' "$value"
}
normalize_issue_ref() {
local ref="$1"
local issue_repo="$REPO"
local issue_num
ref="${ref//$'\r'/}"
ref="${ref//[[:space:]]/}"
if [[ -z "$ref" ]]; then
return 1
fi
if [[ "$ref" == */*#* ]]; then
issue_repo="${ref%#*}"
issue_num="${ref##*#}"
elif [[ "$ref" == [#]* ]]; then
issue_num="${ref#\#}"
else
issue_num="$ref"
fi
if ! [[ "$issue_num" =~ ^[0-9]+$ ]]; then
return 2
fi
printf '%s;%s' "$issue_repo" "$issue_num"
}
REPO="rustfs/rustfs"
WORKFLOWS="ci.yml,e2e-s3tests.yml,e2e-replication-nightly.yml,mint.yml,fuzz.yml"
BRANCH="main"
RUNS=5
REQUIRE_COMPLETE="false"
PR_NUMBER=""
ISSUES=""
while [[ $# -gt 0 ]]; do
case "$1" in
--repo)
REPO="$(arg_value "$1" "${2:-}")"
shift 2
;;
--workflows)
WORKFLOWS="$(arg_value "$1" "${2:-}")"
shift 2
;;
--branch)
BRANCH="$(arg_value "$1" "${2:-}")"
shift 2
;;
--runs)
RUNS="$(arg_value "$1" "${2:-}")"
shift 2
;;
--require-complete)
REQUIRE_COMPLETE="true"
shift
;;
--pr)
PR_NUMBER="$(arg_value "$1" "${2:-}")"
shift 2
;;
--issues)
ISSUES="$(arg_value "$1" "${2:-}")"
shift 2
;;
-h|--help)
usage
exit 0
;;
*)
echo "ERROR: unknown arg: $1" >&2
usage
exit 1
;;
esac
done
require_cmd gh
require_cmd jq
require_cmd date
if ! [[ "${RUNS}" =~ ^[0-9]+$ ]] || (( RUNS == 0 )); then
echo "ERROR: --runs must be a positive integer" >&2
exit 1
fi
run_failed=0
active_found=0
check_workflow() {
local workflow="$1"
local records
local line_count
if ! records="$(gh run list --repo "$REPO" --workflow "$workflow" --branch "$BRANCH" --limit "$RUNS" --json databaseId,name,status,conclusion,headBranch,headSha,startedAt,updatedAt,url | jq -r '.[] | "\(.name)\t\(.status)\t\(.conclusion // "none")\t\(.headBranch // "")\t\(.headSha // "")\t\(.url)"')"; then
echo "ERROR: cannot query workflow=${workflow} in ${REPO}" >&2
run_failed=1
return
fi
if [[ -z "$records" ]]; then
echo "[warn] no runs for workflow=${workflow}" >&2
return
fi
line_count=$(printf '%s\n' "$records" | wc -l | tr -d ' ')
echo "workflow=${workflow} samples=${line_count}/${RUNS}"
while IFS=$'\t' read -r wf status conclusion head_branch head_sha url; do
if [[ "$status" != completed ]]; then
echo " - ${status} | ${conclusion} | branch=${head_branch} | sha=${head_sha:0:9}... | ${url}"
active_found=1
else
echo " - ${status} | ${conclusion} | branch=${head_branch} | sha=${head_sha:0:9}... | ${url}"
fi
if [[ "$status" == "completed" && "$conclusion" != "success" && "$conclusion" != "skipped" && "$conclusion" != "neutral" ]]; then
run_failed=1
fi
done < <(printf '%s\n' "$records")
if [[ "$active_found" == "1" && "$REQUIRE_COMPLETE" == "true" ]]; then
run_failed=1
fi
}
check_pr_checks() {
if [[ -z "$PR_NUMBER" ]]; then
return
fi
if ! records="$(gh pr checks "$PR_NUMBER" --repo "$REPO" --json name,state,conclusion,detailsUrl 2>/dev/null)"; then
echo "ERROR: cannot query PR checks for #${PR_NUMBER} in ${REPO}" >&2
run_failed=1
return
fi
if [[ "$records" == "[]" ]]; then
echo "WARN: PR #${PR_NUMBER} has no recorded checks in ${REPO}"
return
fi
local failed_checks
failed_checks="$(echo "$records" | jq -r '.[] | select(.state != "completed" or (.conclusion != null and (.conclusion != "success" and .conclusion != "neutral" and .conclusion != "skipped"))) | "\(.name)\t\(.state)\t\(.conclusion // \"none\")\t\(.detailsUrl // \"\")"')"
echo "pr=${PR_NUMBER} check-runs"
if [[ -n "$failed_checks" ]]; then
echo "$failed_checks" | while IFS=$'\t' read -r name state conclusion url; do
echo " - ${name}: ${state}/${conclusion} -> ${url}"
done
run_failed=1
else
echo " - all checks passed or pending"
fi
}
check_issues() {
local issues_csv="$1"
local issue_num
local issue_repo
local issue_id
local resolved
local issue_rows
local updated
local issue_data
local state
local title
local url
if [[ -z "$issues_csv" ]]; then
return
fi
IFS=',' read -r -a issue_arr <<< "$issues_csv"
issue_rows=0
echo "## Manual transition follow-up snapshot"
echo "issue_followup_snapshot:"
echo "| issue | title | state | updated_at | link |"
echo "| --- | --- | --- | --- | --- |"
for issue_num in "${issue_arr[@]}"; do
[[ -z "${issue_num}" ]] && continue
if ! resolved="$(normalize_issue_ref "${issue_num}")"; then
echo "| ${issue_num} | invalid format | unknown | unknown | https://github.com/${REPO}/issues/${issue_num} |"
continue
fi
issue_repo="${resolved%;*}"
issue_id="${resolved#*;}"
if ! issue_data="$(gh issue view "$issue_id" --repo "$issue_repo" --json number,title,state,updatedAt,url 2>/dev/null)"; then
echo "| #${issue_id} (${issue_repo}) | unavailable | unknown | unavailable | https://github.com/${issue_repo}/issues/${issue_id} |"
continue
fi
state="$(echo "$issue_data" | jq -r '.state // "unknown"')"
title="$(echo "$issue_data" | jq -r '.title // "unknown"')"
updated="$(echo "$issue_data" | jq -r '.updatedAt // "unknown"')"
url="$(echo "$issue_data" | jq -r '.url // ""')"
if [[ -z "$url" ]]; then
url="https://github.com/${issue_repo}/issues/${issue_id}"
fi
echo "| #${issue_id} (${issue_repo}) | ${title} | ${state} | ${updated} | ${url} |"
issue_rows=$((issue_rows + 1))
done
if (( issue_rows == 0 )); then
echo "| (none) | — | — | — | — |"
fi
}
IFS=',' read -r -a workflow_arr <<< "$WORKFLOWS"
for workflow in "${workflow_arr[@]}"; do
workflow="${workflow//$'\r'/}"
workflow="${workflow//[[:space:]]/}"
[[ -z "$workflow" ]] && continue
check_workflow "$workflow"
done
check_pr_checks
check_issues "$ISSUES"
echo "summary"
echo " - failed_runs: $run_failed"
echo " - active_runs: $active_found"
echo " - repo: $REPO"
echo " - branch: $BRANCH"
echo " - workflows: $WORKFLOWS"
echo " - sampled: $RUNS"
timestamp=$(date -u +%Y-%m-%dT%H:%M:%SZ)
echo " - scanned_at: $timestamp"
if (( run_failed == 1 )); then
exit 1
fi
exit 0
@@ -88,5 +88,29 @@ rg -q "UNKNOWN_FAILURE_RATIO_THRESHOLD" "$TMP_DIR/stress-runbook/run_nightly_str
rg -q "QUEUE_MISMATCH_RATIO_THRESHOLD" "$TMP_DIR/stress-runbook/run_nightly_stress_plan.sh"
rg -q "UNKNOWN_FAILURE_COUNT_THRESHOLD" "$TMP_DIR/stress-runbook/run_nightly_stress_plan.sh"
rg -q "snapshot_failure" "$TMP_DIR/stress-runbook/run_nightly_stress_plan.sh"
rg -q "is_terminal_status" "$TMP_DIR/stress-runbook/run_nightly_stress_plan.sh"
rg -q "status_json" "$TMP_DIR/stress-runbook/run_nightly_stress_plan.sh"
rg -q "unknown_failure_ratio" "$TMP_DIR/stress-runbook/run_nightly_stress_plan.sh"
rg -q "queue_mismatch_ratio" "$TMP_DIR/stress-runbook/run_nightly_stress_plan.sh"
rg -q "report_unknown_failure" "$TMP_DIR/stress-runbook/run_nightly_stress_plan.sh"
rg -q "Manual transition nightly/stress stress-runbook" "$TMP_DIR/stress-runbook/manual_transition_nightly_stress_runbook.md"
rg -q "SOAK_MATRIX_CSV='$TMP_DIR/stress-runbook/nightly_soak_matrix.csv'" "$TMP_DIR/stress-runbook/run_nightly_stress_plan.sh"
bash scripts/monitor_manual_transition_ci.sh --help | rg -q "Usage:"
if bash scripts/monitor_manual_transition_ci.sh --issues >/tmp/monitor_manual_transition_ci.err 2>&1; then
echo "monitor script should fail when --issues has no value" >&2
exit 1
fi
if ! rg -q "ERROR: missing value for --issues" /tmp/monitor_manual_transition_ci.err; then
echo "monitor script missing missing-value guard for --issues" >&2
exit 1
fi
if bash scripts/monitor_manual_transition_ci.sh --runs 0 >/tmp/monitor_manual_transition_ci.err 2>&1; then
echo "monitor script should fail on invalid --runs" >&2
exit 1
fi
if ! rg -q "ERROR: --runs must be a positive integer" /tmp/monitor_manual_transition_ci.err; then
echo "monitor script missing invalid --runs guard output" >&2
exit 1
fi