Files
oa-monitor/collect-business.sh
oa-monitor 20fa48fb34 fix: check-alert v3 修复钉钉推送断档 P0 假阳性
根因:sqlite_exec 吞掉 stderr(2>/dev/null),查询失败返回空串,
0 默认为 0 导致误报

修复:业务层取数显式调用 sqlite3 + 检查退出码 + 正则校验,
查询失败跳过检测并 warn 日志
2026-11-04 19:41:18 +08:00

36 lines
1.4 KiB
Bash

#!/usr/bin/env bash
# ============================================================================
# OA Monitor · 业务指标采集(每 1 小时)
# ============================================================================
set -u
SCRIPT_DIR="$(cd "$(dirname "$0")" && pwd)"
source "${SCRIPT_DIR}/../lib/common.sh"
NOW=$(date +%s)
db_exec() {
local sql="$1"
MYSQL_PWD="${DB_PASS}" timeout 30 mysql -h39.100.72.102 -P3306 -uecology ecology -e "${sql}" 2>/dev/null
}
MSG_TOTAL=$(db_exec "SELECT COUNT(*) FROM ecology_message_log;")
MSG_TOTAL=${MSG_TOTAL:-0}
MSG_FAIL_HOUR=$(db_exec "SELECT COUNT(*) FROM ecology_message_log WHERE retcode != 0 AND operTime >= DATE_SUB(NOW(), INTERVAL 1 HOUR);")
MSG_FAIL_HOUR=${MSG_FAIL_HOUR:-0}
ONGOING=$(db_exec "SELECT COUNT(*) FROM workflow_requestbase WHERE status = 2 OR (status IS NULL AND currentnodeid > 0);")
ONGOING=${ONGOING:-0}
OVERDUE=$(db_exec "SELECT COUNT(*) FROM workflow_requestlog WHERE timeout = 1 AND operatedate >= DATE_SUB(NOW(), INTERVAL 1 DAY);")
OVERDUE=${OVERDUE:-0}
if [[ ${MSG_TOTAL} -gt 0 ]]; then
FAIL_RATE=$(echo "scale=4; ${MSG_FAIL_HOUR} / (${MSG_TOTAL} + 0.001);" | bc 2>/dev/null || echo "0.0")
else
FAIL_RATE="0.0"
fi
sqlite_exec "INSERT INTO business_metrics VALUES (
${NOW}, ${MSG_TOTAL}, ${FAIL_RATE}, ${ONGOING}, ${OVERDUE}
);"
log_info "[BIZ] ✅ msg_total=${MSG_TOTAL} fail_hour=${MSG_FAIL_HOUR} fail_rate=${FAIL_RATE} ongoing=${ONGOING} overdue=${OVERDUE}"