opentsdb操作草稿
插入数据api/put
192.168.1.68:4242/api/put?details
http://localhost:4242/api/put?details
method:POST
[ { "metric": "gas_patrol", "timestamp": 1571028638, "value": 58, "tags": { "host": "web01", "dc": "lga" } }, { "metric": "gas_patrol", "timestamp": 1571028638, "value": 18, "tags": { "host": "web02", "dc": "lga2" } } ]
return
{
"success": 2,
"failed": 0,
"errors": []
}
查询api/query
http://localhost:4242/api/query method:POST { "start": 1571028630, "end": 1571028639, "showTSUIDs":"true", "queries": [ { "aggregator": "avg", "metric": "gas_patrol", "tags": { "host": "web*", "dc": "lga" } } ] }
return
[ { "metric": "gas_patrol", "tags": { "host": "web01", "dc": "lga" }, "aggregateTags": [], "tsuids": [ "000001000001000001000002000002" ], "dps": { "1571028635": 18, "1571028636": 28, "1571028637": 38, "1571028638": 58 } }, { "metric": "gas_patrol", "tags": { "host": "web02", "dc": "lga" }, "aggregateTags": [], "tsuids": [ "000001000001000003000002000002" ], "dps": { "1571028637": 38, "1571028638": 18 } }, { "metric": "gas_patrol", "tags": { "host": "web02", "dc": "lga2" }, "aggregateTags": [], "tsuids": [ "000001000001000003000002000004" ], "dps": { "1571028638": 18 } } ]
平均值:"downsample":"0all-avg"
{ "start": 1271028034, "end": 1571028939, "showTSUIDs":"true", "queries": [ { "aggregator": "avg", "metric": "gas_patrol", "downsample":"0all-avg", "tags": { "host": "web*", "dc": "lga" } } ] }
求和:"downsample":"0all-sum"
下采样(按1分钟/1小时/1天聚合)
"downsample":"1m-avg" "downsample":"1h-avg" "downsample":"1d-avg" "downsample":"0y-avg"
可以查询相应的 5分钟 10小时 22天 等等
下采样补值:
1m-avg-null 1m-avg-zero 。。。 可用的政策包括: None(none) - 默认行为,在序列化期间不输出缺失值,在聚合序列时执行线性插值(或以其他方式指定插值)的。 NaN(nan) - 当序列中缺少所有值时,在序列化输出中输出NaN 。缺少值时跳过聚合中的NaN,而不是将整个分组计算转换为NaN(参见下表)。 Null(null) - 与NaN的行为相同,只是在序列化期间它会发出一个null,而不是一个NaN。 零(zero) - 缺少时间戳时替换为零。零值将合并到汇总结果中。