style: black格式化ci_trigger_monitor.py
CI Build & Deploy Pipeline / Build Staging API Image (pull_request) Has been skipped
CI Build & Deploy Pipeline / Build Staging Web Image (pull_request) Has been skipped
CI Build & Deploy Pipeline / Build Staging Worker Image (pull_request) Has been skipped
CI Build & Deploy Pipeline / Deploy Staging (Watchtower auto-deploy) (pull_request) Has been skipped
CI Build & Deploy Pipeline / Build Production API Image (pull_request) Has been skipped
CI Build & Deploy Pipeline / Build Production Web Image (pull_request) Has been skipped
CI Build & Deploy Pipeline / Build Production Worker Image (pull_request) Has been skipped
CI Build & Deploy Pipeline / Staging E2E Tests (pull_request) Has been skipped
CI Build & Deploy Pipeline / Staging API Integration Tests (pull_request) Has been skipped
CI Build & Deploy Pipeline / Deploy Production (pull_request) Has been skipped
CI Build & Deploy Pipeline / Production Browser E2E (pull_request) Has been skipped
CI/CD Pipeline / Check if frontend-only change (pull_request) Successful in 17s
CI/CD Pipeline / Frontend Lint (pull_request) Successful in 48s
AI Code Review / AI Code Review (pull_request) Successful in 2m15s
CI/CD Pipeline / Validate Code Quality And Tests (pull_request) Failing after 59s
CI/CD Pipeline / Unit Tests (pull_request) Successful in 3m1s
CI/CD Pipeline / Integration Tests (pull_request) Successful in 1m53s

This commit is contained in:
2026-07-17 09:44:38 +08:00
parent 6f1d7f2508
commit 29cb357fc8
+49 -57
View File
@@ -30,23 +30,23 @@ def api_get(path):
token = get_env("GITEA_API_TOKEN")
base_url = get_env("GITEA_URL", "https://git.xiaoxiajianji.com")
repo = get_env("GITEA_REPO", "xiaoxia/xiaoxia-saas")
url = f"{base_url}/api/v1/repos/{repo}{path}"
req = urllib.request.Request(url)
req.add_header("Authorization", f"token {token}")
for attempt in range(3):
try:
with urllib.request.urlopen(req, timeout=30) as resp:
return json.loads(resp.read())
except urllib.error.HTTPError as e:
if e.code >= 500 and attempt < 2:
time.sleep(2 ** attempt)
time.sleep(2**attempt)
continue
raise
except Exception as e:
if attempt < 2:
time.sleep(2 ** attempt)
time.sleep(2**attempt)
continue
raise
@@ -77,18 +77,18 @@ def get_commit_status(sha):
def has_ci_started(statuses):
"""判断是否有CI job已经启动(pending/running/success/failure都算启动了)"""
pr_statuses = [s for s in statuses if 'pull_request' in s.get('context', '')]
pr_statuses = [s for s in statuses if "pull_request" in s.get("context", "")]
if not pr_statuses:
return False
# 只要有非pending且非空的状态,就算启动了
for s in pr_statuses:
if s.get('status') in ['success', 'failure', 'running']:
if s.get("status") in ["success", "failure", "running"]:
return True
if s.get('status') == 'pending' and 'Has started running' in s.get('description', ''):
if s.get("status") == "pending" and "Has started running" in s.get("description", ""):
return True
# 全是"Blocked by required conditions"的pending也算(说明CI系统收到了事件)
for s in pr_statuses:
if 'Blocked' in s.get('description', ''):
if "Blocked" in s.get("description", ""):
return True
return False
@@ -99,26 +99,23 @@ def send_alert(pr_num, pr_title, pr_url, head_sha, commit_age_min):
if not webhook:
print(" ⚠️ 未配置CI_NOTIFY_WEBHOOK,跳过告警")
return
gitea_url = get_env("GITEA_URL", "https://git.xiaoxiajianji.com")
content = {
"msg_type": "interactive",
"card": {
"header": {
"title": {
"tag": "plain_text",
"content": f"⚠️ CI告警 - PR#{pr_num} CI未触发"
},
"template": "red"
"title": {"tag": "plain_text", "content": f"⚠️ CI告警 - PR#{pr_num} CI未触发"},
"template": "red",
},
"elements": [
{
"tag": "div",
"text": {
"tag": "lark_md",
"content": f"**PR**: [{pr_title}]({pr_url})\n**最新commit**: `{head_sha[:12]}`\n**已等待**: {commit_age_min:.0f} 分钟仍无CI启动\n**可能原因**: Gitea Actions事件丢失 / Webhook失败 / Runner资源不足"
}
"content": f"**PR**: [{pr_title}]({pr_url})\n**最新commit**: `{head_sha[:12]}`\n**已等待**: {commit_age_min:.0f} 分钟仍无CI启动\n**可能原因**: Gitea Actions事件丢失 / Webhook失败 / Runner资源不足",
},
},
{
"tag": "action",
@@ -127,29 +124,29 @@ def send_alert(pr_num, pr_title, pr_url, head_sha, commit_age_min):
"tag": "button",
"text": {"tag": "plain_text", "content": "查看PR"},
"url": pr_url,
"type": "primary"
"type": "primary",
},
{
"tag": "button",
"text": {"tag": "plain_text", "content": "查看Actions"},
"url": f"{pr_url}/files",
"type": "default"
}
]
"type": "default",
},
],
},
{
"tag": "note",
"elements": [
{"tag": "plain_text", "content": f"CI触发监控 | 检测时间: {time.strftime('%Y-%m-%d %H:%M:%S')}"}
]
}
]
}
],
},
],
},
}
try:
data = json.dumps(content).encode()
req = urllib.request.Request(webhook, data=data, method='POST')
req = urllib.request.Request(webhook, data=data, method="POST")
req.add_header("Content-Type", "application/json")
with urllib.request.urlopen(req, timeout=10) as resp:
resp.read()
@@ -160,86 +157,81 @@ def send_alert(pr_num, pr_title, pr_url, head_sha, commit_age_min):
def main():
stale_threshold = int(get_env("STALE_THRESHOLD_MIN", "5"))
print("=" * 60)
print(f"CI触发监控 - 检测时间: {time.strftime('%Y-%m-%d %H:%M:%S')}")
print(f"告警阈值: {stale_threshold}分钟无CI启动")
print("=" * 60)
# 获取open PR列表
try:
prs = get_open_prs()
except Exception as e:
print(f"❌ 获取PR列表失败: {e}")
sys.exit(0) # 告警脚本不阻断CI
print(f"\n{len(prs)} 个open PR\n")
stale_prs = []
now = time.time()
for pr in prs:
pr_num = pr['number']
pr_title = pr['title']
pr_url = pr['html_url']
head_sha = pr['head']['sha']
updated_at = pr['updated_at']
pr_num = pr["number"]
pr_title = pr["title"]
pr_url = pr["html_url"]
head_sha = pr["head"]["sha"]
updated_at = pr["updated_at"]
# 解析updated_atISO格式)
try:
# 2026-07-17T09:22:43+08:00
from datetime import datetime, timezone, timedelta
# 简化处理:直接用字符串解析
ts_str = updated_at.replace('Z', '+00:00')
ts_str = updated_at.replace("Z", "+00:00")
# 手动解析
dt = datetime.fromisoformat(ts_str)
commit_time = dt.timestamp()
except Exception as e:
print(f" ⚠️ PR#{pr_num} 时间解析失败: {e}")
continue
age_min = (now - commit_time) / 60
print(f"PR#{pr_num:3d} | {pr_title[:45]:45s} | 更新于 {age_min:.0f}min前")
# 少于2分钟的跳过,给CI一点启动时间
if age_min < 2:
print(f" ⏳ 刚更新,等待CI启动...")
continue
# 获取commit状态
status = get_commit_status(head_sha)
statuses = status.get('statuses', [])
statuses = status.get("statuses", [])
if has_ci_started(statuses):
print(f" ✅ CI已启动 (state={status.get('state')})")
continue
# CI未启动,判断是否超过阈值
if age_min >= stale_threshold:
print(f" 🚨 CI未触发!已等待 {age_min:.0f} 分钟")
stale_prs.append({
'num': pr_num,
'title': pr_title,
'url': pr_url,
'sha': head_sha,
'age_min': age_min
})
stale_prs.append({"num": pr_num, "title": pr_title, "url": pr_url, "sha": head_sha, "age_min": age_min})
else:
print(f" ⏳ CI尚未启动 ({age_min:.0f}min < {stale_threshold}min阈值)")
# 发送告警
print(f"\n{'=' * 60}")
print(f"检测结果: {len(stale_prs)} 个PR CI未触发超过阈值")
if stale_prs:
print("\n告警列表:")
for pr in stale_prs:
print(f" - PR#{pr['num']}: {pr['title'][:40]} ({pr['age_min']:.0f}min)")
send_alert(pr['num'], pr['title'], pr['url'], pr['sha'], pr['age_min'])
send_alert(pr["num"], pr["title"], pr["url"], pr["sha"], pr["age_min"])
else:
print("✅ 所有PR CI触发正常")
print("=" * 60)