trying use approximate histogram, get a problem of lower limit and upper limit

i am trying use approxHistogramFold, I set the lower Limit and uppder Limit when ingest data,

below is part of the ingest task log:

“metricsSpec” : [ {

“type” : “longSum”,

“name” : “total_num”,

“fieldName” : “total_num”

}, {

“type” : “doubleSum”,

“name” : “total_time”,

“fieldName” : “total_time”

}, {

“type” : “doubleSum”,

“name” : “byteSend”,

“fieldName” : “byteSend”

}, {

“type” : “doubleSum”,

“name” : “byteReceived”,

“fieldName” : “byteReceived”

}, {

“type” : “approxHistogramFold”,

    "name" : "total_time_histogram",

    "fieldName" : "total_time",

    "resolution" : 500,

    "numBuckets" : 500,

    "lowerLimit" : 0.0,

    "upperLimit" : 10.0

} ],

but when I am querying using this query:

{

“queryType”:“timeseries”,

“dataSource”:“httptransaction500”,

“threshold”: 5,

“dimensions”:[

{ “type” : “default”, “dimension” : “domain_id”, “outputName”: “domain_id” }

],

“granularity”: {“type”: “duration”, “duration”: 600000, “origin”: “2016-02-16T00:00:00Z”},

“filter”:{

“type”: “selector”,

“dimension”: “app_id”,

“value”: “9084”

},

“aggregations”: [

{

“type”: “longSum”,

“name”: “total_num”,

“fieldName”: “total_num”

},

{

“type”: “approxHistogramFold”,

“name”: “total_time_histogram”,

“fieldName”: “total_time_histogram”,

“lowerLimit” : 0,

“upperLimit” : 10

}

],

“postAggregations”:[

{ “type” : “quantiles”, “name” : “sla”, “fieldName” : “total_time_histogram”,“probabilities” : [ 0.5,0.75,0.9,0.95,0.99] }

],

“intervals”: [

“2016-02-16T00:00:00.000/2016-02-17T00:00:00.000”

]

}

druid give result:

{

“timestamp” : “2016-02-16T10:50:00.000Z”,

“result” : {

“total_num” : 21923,

“sla” : {

“probabilities” : [ 0.5, 0.75, 0.9, 0.95, 0.99 ],

“quantiles” : [ 2.0E-5, 0.576681, 1.6588136, 2.670443, 27.418663 ],

“min” : 2.0E-5,

“max” : 497.55

},

“total_time_histogram” : {

“breaks” : [ -82.92497253417969, 2.288818359375E-5, 82.92501831054688, 165.85000610351562, 248.77499389648438, 331.6999816894531, 414.6249694824219, 497.54998779296875 ],

“counts” : [ 1.0658757219061954E-6, 21813.169921875, 39.539031982421875, 30.752580642700195, 21.966129302978516, 13.179678916931152, 4.393228530883789 ]

}

}

}

still have more than one breaks which contains the upper limit(10), and break [ 2.288818359375E-5, 82.92501831054688] almost take 99% values which cause a bad result.

it seems that the lower and upper limit, it does not work in some way. is there some thing I config wrong?

thx.

Hey 何文斌,

I took a look at the code and I think the objects are getting created properly, but the folding might be doing something unexpected. Perhaps if you posted a sample of data that could reproduce this issue, that would help figure out what is going on.

hey, Gian
I’ve upload a sample file, please help, thx.

在 2016年2月17日星期三 UTC+8上午11:41:15,Gian Merlino写道:

log-http-1w.txt (1.57 MB)

is there any body can help??? thx

在 2016年3月10日星期四 UTC+8下午1:54:26,何文斌写道:

It takes us a few days to get through all the community emails. We can try to look into this in the next few days. Also, if you need dedicated support, check out http://imply.io/

thx

在 2016年3月17日星期四 UTC+8下午2:21:28,Fangjin Yang写道: