187 lines
7.9 KiB
Go
187 lines
7.9 KiB
Go
|
package v3
|
||
|
|
||
|
import (
|
||
|
"fmt"
|
||
|
"math"
|
||
|
"strings"
|
||
|
|
||
|
"go.signoz.io/signoz/pkg/query-service/app/metrics/v4/helpers"
|
||
|
"go.signoz.io/signoz/pkg/query-service/constants"
|
||
|
v3 "go.signoz.io/signoz/pkg/query-service/model/v3"
|
||
|
"go.signoz.io/signoz/pkg/query-service/utils"
|
||
|
)
|
||
|
|
||
|
// This logic is little convoluted for a reason.
|
||
|
// When we work with cumulative metrics, the table view need to show the data for the entire time range.
|
||
|
// In some cases, we could take the points at the start and end of the time range and divide it by the
|
||
|
// duration. But, the problem is there is no guarantee that the trend will be linear between the start and end.
|
||
|
// We can sum the rate of change for some interval X, this interval can be step size of time series.
|
||
|
// However, the speed of query depends on the number of timestamps, so we bump up the xx the step size.
|
||
|
// This should be a good balance between speed and accuracy.
|
||
|
// TODO: find a better way to do this
|
||
|
func stepForTableCumulative(start, end int64) int64 {
|
||
|
// round up to the nearest multiple of 60
|
||
|
duration := (end - start + 1) / 1000
|
||
|
step := math.Max(math.Floor(float64(duration)/120), 60) // assuming 120 max points
|
||
|
if duration > 1800 { // bump for longer duration
|
||
|
step = step * 5
|
||
|
}
|
||
|
return int64(step)
|
||
|
}
|
||
|
|
||
|
func buildMetricQueryForTable(start, end, _ int64, mq *v3.BuilderQuery) (string, error) {
|
||
|
|
||
|
step := stepForTableCumulative(start, end)
|
||
|
|
||
|
points := ((end - start + 1) / 1000) / step
|
||
|
|
||
|
metricQueryGroupBy := mq.GroupBy
|
||
|
|
||
|
filterSubQuery, err := helpers.PrepareTimeseriesFilterQueryV3(start, end, mq)
|
||
|
if err != nil {
|
||
|
return "", err
|
||
|
}
|
||
|
|
||
|
samplesTableTimeFilter := fmt.Sprintf("metric_name = %s AND unix_milli >= %d AND unix_milli <= %d", utils.ClickHouseFormattedValue(mq.AggregateAttribute.Key), start, end)
|
||
|
|
||
|
// Select the aggregate value for interval
|
||
|
queryTmplCounterInner :=
|
||
|
"SELECT %s" +
|
||
|
" toStartOfInterval(toDateTime(intDiv(unix_milli, 1000)), INTERVAL %d SECOND) as ts," +
|
||
|
" %s as value" +
|
||
|
" FROM " + constants.SIGNOZ_METRIC_DBNAME + "." + constants.SIGNOZ_SAMPLES_V4_TABLENAME +
|
||
|
" INNER JOIN" +
|
||
|
" (%s) as filtered_time_series" +
|
||
|
" USING fingerprint" +
|
||
|
" WHERE " + samplesTableTimeFilter +
|
||
|
" GROUP BY %s" +
|
||
|
" ORDER BY %s ts"
|
||
|
|
||
|
// Select the aggregate value for interval
|
||
|
queryTmpl :=
|
||
|
"SELECT %s" +
|
||
|
" toStartOfHour(now()) as ts," + // now() has no menaing & used as a placeholder for ts
|
||
|
" %s as value" +
|
||
|
" FROM " + constants.SIGNOZ_METRIC_DBNAME + "." + constants.SIGNOZ_SAMPLES_V4_TABLENAME +
|
||
|
" INNER JOIN" +
|
||
|
" (%s) as filtered_time_series" +
|
||
|
" USING fingerprint" +
|
||
|
" WHERE " + samplesTableTimeFilter +
|
||
|
" GROUP BY %s" +
|
||
|
" ORDER BY %s ts"
|
||
|
|
||
|
// tagsWithoutLe is used to group by all tags except le
|
||
|
// This is done because we want to group by le only when we are calculating quantile
|
||
|
// Otherwise, we want to group by all tags except le
|
||
|
tagsWithoutLe := []string{}
|
||
|
for _, tag := range mq.GroupBy {
|
||
|
if tag.Key != "le" {
|
||
|
tagsWithoutLe = append(tagsWithoutLe, tag.Key)
|
||
|
}
|
||
|
}
|
||
|
|
||
|
// orderWithoutLe := orderBy(mq.OrderBy, tagsWithoutLe)
|
||
|
|
||
|
groupByWithoutLe := groupBy(tagsWithoutLe...)
|
||
|
groupTagsWithoutLe := groupSelect(tagsWithoutLe...)
|
||
|
orderWithoutLe := orderBy(mq.OrderBy, tagsWithoutLe)
|
||
|
|
||
|
groupBy := groupByAttributeKeyTags(metricQueryGroupBy...)
|
||
|
groupTags := groupSelectAttributeKeyTags(metricQueryGroupBy...)
|
||
|
orderBy := orderByAttributeKeyTags(mq.OrderBy, metricQueryGroupBy)
|
||
|
|
||
|
if len(orderBy) != 0 {
|
||
|
orderBy += ","
|
||
|
}
|
||
|
if len(orderWithoutLe) != 0 {
|
||
|
orderWithoutLe += ","
|
||
|
}
|
||
|
|
||
|
switch mq.AggregateOperator {
|
||
|
case v3.AggregateOperatorRate:
|
||
|
return "", fmt.Errorf("rate is not supported for table view")
|
||
|
case v3.AggregateOperatorSumRate, v3.AggregateOperatorAvgRate, v3.AggregateOperatorMaxRate, v3.AggregateOperatorMinRate:
|
||
|
rateGroupBy := "fingerprint, " + groupBy
|
||
|
rateGroupTags := "fingerprint, " + groupTags
|
||
|
rateOrderBy := "fingerprint, " + orderBy
|
||
|
partitionBy := "fingerprint"
|
||
|
if len(groupTags) != 0 {
|
||
|
partitionBy += ", " + groupTags
|
||
|
partitionBy = strings.Trim(partitionBy, ", ")
|
||
|
}
|
||
|
op := "max(value)"
|
||
|
subQuery := fmt.Sprintf(
|
||
|
queryTmplCounterInner, rateGroupTags, step, op, filterSubQuery, rateGroupBy, rateOrderBy,
|
||
|
) // labels will be same so any should be fine
|
||
|
query := `SELECT %s ts, ` + rateWithoutNegative + `as rate_value FROM(%s) WINDOW rate_window as (PARTITION BY %s ORDER BY %s ts)`
|
||
|
query = fmt.Sprintf(query, groupTags, subQuery, partitionBy, rateOrderBy)
|
||
|
query = fmt.Sprintf(`SELECT %s toStartOfHour(now()) as ts, %s(rate_value)/%d as value FROM (%s) WHERE isNaN(rate_value) = 0 GROUP BY %s ORDER BY %s ts`, groupTags, aggregateOperatorToSQLFunc[mq.AggregateOperator], points, query, groupBy, orderBy)
|
||
|
return query, nil
|
||
|
case
|
||
|
v3.AggregateOperatorRateSum,
|
||
|
v3.AggregateOperatorRateMax,
|
||
|
v3.AggregateOperatorRateAvg,
|
||
|
v3.AggregateOperatorRateMin:
|
||
|
step = ((end - start + 1) / 1000) / 2
|
||
|
op := fmt.Sprintf("%s(value)", aggregateOperatorToSQLFunc[mq.AggregateOperator])
|
||
|
subQuery := fmt.Sprintf(queryTmplCounterInner, groupTags, step, op, filterSubQuery, groupBy, orderBy)
|
||
|
partitionBy := ""
|
||
|
if len(groupTags) != 0 {
|
||
|
partitionBy = "PARTITION BY " + groupTags
|
||
|
partitionBy = strings.Trim(partitionBy, ", ")
|
||
|
}
|
||
|
query := `SELECT %s toStartOfHour(now()) as ts, ` + rateWithoutNegative + `as value FROM(%s) WINDOW rate_window as (%s ORDER BY %s ts)`
|
||
|
query = fmt.Sprintf(query, groupTags, subQuery, partitionBy, groupTags)
|
||
|
return query, nil
|
||
|
case
|
||
|
v3.AggregateOperatorP05,
|
||
|
v3.AggregateOperatorP10,
|
||
|
v3.AggregateOperatorP20,
|
||
|
v3.AggregateOperatorP25,
|
||
|
v3.AggregateOperatorP50,
|
||
|
v3.AggregateOperatorP75,
|
||
|
v3.AggregateOperatorP90,
|
||
|
v3.AggregateOperatorP95,
|
||
|
v3.AggregateOperatorP99:
|
||
|
op := fmt.Sprintf("quantile(%v)(value)", aggregateOperatorToPercentile[mq.AggregateOperator])
|
||
|
query := fmt.Sprintf(queryTmpl, groupTags, op, filterSubQuery, groupBy, orderBy)
|
||
|
return query, nil
|
||
|
case v3.AggregateOperatorHistQuant50, v3.AggregateOperatorHistQuant75, v3.AggregateOperatorHistQuant90, v3.AggregateOperatorHistQuant95, v3.AggregateOperatorHistQuant99:
|
||
|
rateGroupBy := "fingerprint, " + groupBy
|
||
|
rateGroupTags := "fingerprint, " + groupTags
|
||
|
rateOrderBy := "fingerprint, " + orderBy
|
||
|
partitionBy := "fingerprint"
|
||
|
if len(groupTags) != 0 {
|
||
|
partitionBy += ", " + groupTags
|
||
|
partitionBy = strings.Trim(partitionBy, ", ")
|
||
|
}
|
||
|
op := "max(value)"
|
||
|
subQuery := fmt.Sprintf(
|
||
|
queryTmplCounterInner, rateGroupTags, step, op, filterSubQuery, rateGroupBy, rateOrderBy,
|
||
|
) // labels will be same so any should be fine
|
||
|
query := `SELECT %s ts, ` + rateWithoutNegative + ` as rate_value FROM(%s) WINDOW rate_window as (PARTITION BY %s ORDER BY %s ts)`
|
||
|
query = fmt.Sprintf(query, groupTags, subQuery, partitionBy, rateOrderBy)
|
||
|
query = fmt.Sprintf(`SELECT %s toStartOfHour(now()) as ts, sum(rate_value)/%d as value FROM (%s) WHERE isNaN(rate_value) = 0 GROUP BY %s ORDER BY %s ts`, groupTags, points, query, groupBy, orderBy)
|
||
|
value := aggregateOperatorToPercentile[mq.AggregateOperator]
|
||
|
|
||
|
query = fmt.Sprintf(`SELECT %s toStartOfHour(now()) as ts, histogramQuantile(arrayMap(x -> toFloat64(x), groupArray(le)), groupArray(value), %.3f) as value FROM (%s) GROUP BY %s ORDER BY %s ts`, groupTagsWithoutLe, value, query, groupByWithoutLe, orderWithoutLe)
|
||
|
return query, nil
|
||
|
case v3.AggregateOperatorAvg, v3.AggregateOperatorSum, v3.AggregateOperatorMin, v3.AggregateOperatorMax:
|
||
|
op := fmt.Sprintf("%s(value)", aggregateOperatorToSQLFunc[mq.AggregateOperator])
|
||
|
query := fmt.Sprintf(queryTmpl, groupTags, op, filterSubQuery, groupBy, orderBy)
|
||
|
return query, nil
|
||
|
case v3.AggregateOperatorCount:
|
||
|
op := "toFloat64(count(*))"
|
||
|
query := fmt.Sprintf(queryTmpl, groupTags, op, filterSubQuery, groupBy, orderBy)
|
||
|
return query, nil
|
||
|
case v3.AggregateOperatorCountDistinct:
|
||
|
op := "toFloat64(count(distinct(value)))"
|
||
|
query := fmt.Sprintf(queryTmpl, groupTags, op, filterSubQuery, groupBy, orderBy)
|
||
|
return query, nil
|
||
|
case v3.AggregateOperatorNoOp:
|
||
|
return "", fmt.Errorf("noop is not supported for table view")
|
||
|
default:
|
||
|
return "", fmt.Errorf("unsupported aggregate operator")
|
||
|
}
|
||
|
}
|