182 lines
4.6 KiB
JSON
182 lines
4.6 KiB
JSON
[
|
|
{
|
|
"name": "CPU利用率高",
|
|
"note": "",
|
|
"severity": 2,
|
|
"disabled": 0,
|
|
"prom_for_duration": 60,
|
|
"prom_ql": "100 * sum by (instance) (rate(windows_cpu_time_total{mode != 'idle'}[5m])) / count by (instance) (windows_cpu_core_frequency_mhz) > 80",
|
|
"prom_eval_interval": 15,
|
|
"enable_stime": "00:00",
|
|
"enable_etime": "23:59",
|
|
"enable_days_of_week": [
|
|
"1",
|
|
"2",
|
|
"3",
|
|
"4",
|
|
"5",
|
|
"6",
|
|
"0"
|
|
],
|
|
"enable_in_bg": 0,
|
|
"notify_recovered": 1,
|
|
"notify_channels": [],
|
|
"notify_repeat_step": 60,
|
|
"recover_duration": 0,
|
|
"callbacks": [],
|
|
"runbook_url": "",
|
|
"append_tags": [
|
|
"alertname=HighCPUUsage"
|
|
]
|
|
},
|
|
{
|
|
"name": "机器在15分钟内发生过重启",
|
|
"note": "",
|
|
"severity": 3,
|
|
"disabled": 0,
|
|
"prom_for_duration": 60,
|
|
"prom_ql": "time() - windows_system_system_up_time < 900",
|
|
"prom_eval_interval": 15,
|
|
"enable_stime": "00:00",
|
|
"enable_etime": "23:59",
|
|
"enable_days_of_week": [
|
|
"1",
|
|
"2",
|
|
"3",
|
|
"4",
|
|
"5",
|
|
"6",
|
|
"0"
|
|
],
|
|
"enable_in_bg": 0,
|
|
"notify_recovered": 1,
|
|
"notify_channels": [],
|
|
"notify_repeat_step": 60,
|
|
"recover_duration": 0,
|
|
"callbacks": [],
|
|
"runbook_url": "",
|
|
"append_tags": [
|
|
"alertname=UpTimeLessThan15Min"
|
|
]
|
|
},
|
|
{
|
|
"name": "物理内存使用率过高",
|
|
"note": "",
|
|
"severity": 2,
|
|
"disabled": 0,
|
|
"prom_for_duration": 60,
|
|
"prom_ql": "100 * (windows_cs_physical_memory_bytes - windows_os_physical_memory_free_bytes) / windows_cs_physical_memory_bytes > 80",
|
|
"prom_eval_interval": 15,
|
|
"enable_stime": "00:00",
|
|
"enable_etime": "23:59",
|
|
"enable_days_of_week": [
|
|
"1",
|
|
"2",
|
|
"3",
|
|
"4",
|
|
"5",
|
|
"6",
|
|
"0"
|
|
],
|
|
"enable_in_bg": 0,
|
|
"notify_recovered": 1,
|
|
"notify_channels": [],
|
|
"notify_repeat_step": 60,
|
|
"recover_duration": 0,
|
|
"callbacks": [],
|
|
"runbook_url": "",
|
|
"append_tags": [
|
|
"alertname=HighPhysicalMemoryUsage"
|
|
]
|
|
},
|
|
{
|
|
"name": "硬盘快要写满了，注意",
|
|
"note": "",
|
|
"severity": 1,
|
|
"disabled": 0,
|
|
"prom_for_duration": 60,
|
|
"prom_ql": "100 * (windows_logical_disk_size_bytes - windows_logical_disk_free_bytes) / windows_logical_disk_size_bytes > 90",
|
|
"prom_eval_interval": 15,
|
|
"enable_stime": "00:00",
|
|
"enable_etime": "23:59",
|
|
"enable_days_of_week": [
|
|
"1",
|
|
"2",
|
|
"3",
|
|
"4",
|
|
"5",
|
|
"6",
|
|
"0"
|
|
],
|
|
"enable_in_bg": 0,
|
|
"notify_recovered": 1,
|
|
"notify_channels": [],
|
|
"notify_repeat_step": 60,
|
|
"recover_duration": 0,
|
|
"callbacks": [],
|
|
"runbook_url": "",
|
|
"append_tags": [
|
|
"alertname=LogicalDiskFull"
|
|
]
|
|
},
|
|
{
|
|
"name": "系统有In方向丢包情况出现",
|
|
"note": "",
|
|
"severity": 2,
|
|
"disabled": 0,
|
|
"prom_for_duration": 60,
|
|
"prom_ql": "100 * rate(windows_net_packets_received_errors[5m]) / (rate(windows_net_packets_received_errors[5m]) + rate(windows_net_packets_received_total[5m])>0) > 0",
|
|
"prom_eval_interval": 15,
|
|
"enable_stime": "00:00",
|
|
"enable_etime": "23:59",
|
|
"enable_days_of_week": [
|
|
"1",
|
|
"2",
|
|
"3",
|
|
"4",
|
|
"5",
|
|
"6",
|
|
"0"
|
|
],
|
|
"enable_in_bg": 0,
|
|
"notify_recovered": 1,
|
|
"notify_channels": [],
|
|
"notify_repeat_step": 60,
|
|
"recover_duration": 0,
|
|
"callbacks": [],
|
|
"runbook_url": "",
|
|
"append_tags": [
|
|
"alertname=HighInboundErrorRate"
|
|
]
|
|
},
|
|
{
|
|
"name": "系统有Out方向丢包情况出现",
|
|
"note": "",
|
|
"severity": 2,
|
|
"disabled": 0,
|
|
"prom_for_duration": 60,
|
|
"prom_ql": "100 * rate(windows_net_packets_outbound_errors[5m]) / (rate(windows_net_packets_outbound_errors[5m]) + rate(windows_net_packets_sent_total[5m])>0) > 0",
|
|
"prom_eval_interval": 15,
|
|
"enable_stime": "00:00",
|
|
"enable_etime": "23:59",
|
|
"enable_days_of_week": [
|
|
"1",
|
|
"2",
|
|
"3",
|
|
"4",
|
|
"5",
|
|
"6",
|
|
"0"
|
|
],
|
|
"enable_in_bg": 0,
|
|
"notify_recovered": 1,
|
|
"notify_channels": [],
|
|
"notify_repeat_step": 60,
|
|
"recover_duration": 0,
|
|
"callbacks": [],
|
|
"runbook_url": "",
|
|
"append_tags": [
|
|
"alertname=HighOutboundErrorRate"
|
|
]
|
|
}
|
|
] |