mirror of
https://github.com/grafana/tempo.git
synced 2025-07-11 16:59:09 +00:00
895f1872dc
* enforce max series for metrics queries * changelog * manifest.md * bail as soon as max series is hit even if each series does not contain all data points * added test and removed v2 * add querier implementation * lint * enforce limit at metrics evaluator * typo * rebase mishap * lint * lint * fix race maybe * i love racing * lint * handle diff * address comments p1 * return seriesCount instead of using results() * remove channel and break as soon as any layer reaches limit * rebase * make generator series count simpler * add length() function instead * update benchmark test * lint * another lint * atomic should only keep track of raw results * remove max series logic out of handler, add length() to metrics eval * lint * add atomic bool and fix test * oooops * tidy up * rebase and add additional logging for max series * move maxSeries logic * lint * update tests for topk bottomk
1076 lines
36 KiB
Go
1076 lines
36 KiB
Go
package tempodb
|
|
|
|
import (
|
|
"context"
|
|
"math"
|
|
"path"
|
|
"sort"
|
|
"strings"
|
|
"testing"
|
|
"time"
|
|
|
|
"github.com/go-kit/log"
|
|
"github.com/google/go-cmp/cmp"
|
|
"github.com/grafana/tempo/pkg/model"
|
|
"github.com/grafana/tempo/pkg/tempopb"
|
|
common_v1 "github.com/grafana/tempo/pkg/tempopb/common/v1"
|
|
resource_v1 "github.com/grafana/tempo/pkg/tempopb/resource/v1"
|
|
v1 "github.com/grafana/tempo/pkg/tempopb/trace/v1"
|
|
"github.com/grafana/tempo/pkg/traceql"
|
|
"github.com/grafana/tempo/pkg/util/test"
|
|
"github.com/grafana/tempo/tempodb/backend"
|
|
"github.com/grafana/tempo/tempodb/backend/local"
|
|
"github.com/grafana/tempo/tempodb/encoding/common"
|
|
"github.com/grafana/tempo/tempodb/encoding/vparquet4"
|
|
"github.com/grafana/tempo/tempodb/wal"
|
|
"github.com/stretchr/testify/require"
|
|
)
|
|
|
|
func requestWithDefaultRange(q string) *tempopb.QueryRangeRequest {
|
|
return &tempopb.QueryRangeRequest{
|
|
Start: 1,
|
|
End: 50 * uint64(time.Second),
|
|
Step: 15 * uint64(time.Second),
|
|
Query: q,
|
|
}
|
|
}
|
|
|
|
var queryRangeTestCases = []struct {
|
|
name string
|
|
req *tempopb.QueryRangeRequest
|
|
// expectedL1 is the expected result of the first level of aggregation
|
|
expectedL1 []*tempopb.TimeSeries
|
|
// expectedL2 is the expected result of the second level of aggregation
|
|
// if nil, the data is not changed and expected results are from previous step
|
|
expectedL2 []*tempopb.TimeSeries
|
|
// expectedL3 is the expected result of the third level of aggregation
|
|
// if nil, the data is not changed and expected results are from previous step
|
|
expectedL3 []*tempopb.TimeSeries
|
|
}{
|
|
{
|
|
name: "rate",
|
|
req: requestWithDefaultRange("{ } | rate()"),
|
|
expectedL1: []*tempopb.TimeSeries{
|
|
{
|
|
PromLabels: `{__name__="rate"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueString("__name__", "rate")},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 14.0 / 15.0}, // First interval starts at 1, so it only has 14 spans
|
|
{TimestampMs: 15_000, Value: 1.0}, // Spans every 1 second
|
|
{TimestampMs: 30_000, Value: 1.0}, // Spans every 1 second
|
|
{TimestampMs: 45_000, Value: 5.0 / 15.0}, // Interval [45,50) has 5 spans
|
|
{TimestampMs: 60_000, Value: 0}, // I think this is a bug that we extend out an extra interval
|
|
},
|
|
},
|
|
},
|
|
expectedL2: []*tempopb.TimeSeries{
|
|
{
|
|
PromLabels: `{__name__="rate"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueString("__name__", "rate")},
|
|
// with two sources rate will be doubled
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 2 * 14.0 / 15.0},
|
|
{TimestampMs: 15_000, Value: 2 * 1.0},
|
|
{TimestampMs: 30_000, Value: 2 * 1.0},
|
|
{TimestampMs: 45_000, Value: 2 * 5.0 / 15.0},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
},
|
|
},
|
|
{
|
|
name: "rate_with_filter",
|
|
req: requestWithDefaultRange(`{ .service.name="even" } | rate()`),
|
|
expectedL1: []*tempopb.TimeSeries{
|
|
{
|
|
PromLabels: `{__name__="rate"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueString("__name__", "rate")},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 7.0 / 15.0}, // Interval [ 1, 14], 7 spans at 2, 4, 6, 8, 10, 12, 14
|
|
{TimestampMs: 15_000, Value: 7.0 / 15.0}, // Interval [15, 29], 7 spans at 16, 18, 20, 22, 24, 26, 28
|
|
{TimestampMs: 30_000, Value: 8.0 / 15.0}, // Interval [30, 44], 8 spans at 30, 32, 34, 36, 38, 40, 42, 44
|
|
{TimestampMs: 45_000, Value: 2.0 / 15.0}, // Interval [45, 50), 2 spans at 46, 48
|
|
{TimestampMs: 60_000, Value: 0}, // I think this is a bug that we extend out an extra interval
|
|
},
|
|
},
|
|
},
|
|
expectedL2: []*tempopb.TimeSeries{
|
|
{
|
|
PromLabels: `{__name__="rate"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueString("__name__", "rate")},
|
|
// with two sources rate will be doubled
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 2 * 7.0 / 15.0},
|
|
{TimestampMs: 15_000, Value: 2 * 7.0 / 15.0},
|
|
{TimestampMs: 30_000, Value: 2 * 8.0 / 15.0},
|
|
{TimestampMs: 45_000, Value: 2 * 2.0 / 15.0},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
},
|
|
},
|
|
{
|
|
name: "rate_no_spans",
|
|
req: requestWithDefaultRange(`{ .service.name="does_not_exist" } | rate()`),
|
|
expectedL1: []*tempopb.TimeSeries{
|
|
{
|
|
PromLabels: `{__name__="rate"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueString("__name__", "rate")},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 0},
|
|
{TimestampMs: 15_000, Value: 0},
|
|
{TimestampMs: 30_000, Value: 0},
|
|
{TimestampMs: 45_000, Value: 0},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
},
|
|
},
|
|
{
|
|
name: "count_over_time",
|
|
req: requestWithDefaultRange(`{ } | count_over_time()`),
|
|
expectedL1: []*tempopb.TimeSeries{
|
|
{
|
|
PromLabels: `{__name__="count_over_time"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueString("__name__", "count_over_time")},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 14}, // Interval [1, 14], 14 spans
|
|
{TimestampMs: 15_000, Value: 15}, // Interval [15, 29], 15 spans
|
|
{TimestampMs: 30_000, Value: 15}, // Interval [30, 44], 15 spans
|
|
{TimestampMs: 45_000, Value: 5}, // Interval [45, 50), 5 spans
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
},
|
|
expectedL2: []*tempopb.TimeSeries{
|
|
{
|
|
PromLabels: `{__name__="count_over_time"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueString("__name__", "count_over_time")},
|
|
// with two sources count will be doubled
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 2 * 14},
|
|
{TimestampMs: 15_000, Value: 2 * 15},
|
|
{TimestampMs: 30_000, Value: 2 * 15},
|
|
{TimestampMs: 45_000, Value: 2 * 5},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
},
|
|
},
|
|
{
|
|
name: "count_over_time",
|
|
req: requestWithDefaultRange(`{ } | count_over_time() by (.service.name)`),
|
|
expectedL1: []*tempopb.TimeSeries{
|
|
{
|
|
PromLabels: `{".service.name"="even"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueString(".service.name", "even")},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 7}, // [2, 4, 6, 8, 10, 12, 14] - total: 7
|
|
{TimestampMs: 15_000, Value: 7}, // [16, 18, 20, 22, 24, 26, 28] - total: 7
|
|
{TimestampMs: 30_000, Value: 8}, // [30, 32, 34, 36, 38, 40, 42, 44] - total: 8
|
|
{TimestampMs: 45_000, Value: 2}, // [46, 48]
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
{
|
|
PromLabels: `{".service.name"="odd"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueString(".service.name", "odd")},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 7}, // [1, 3, 5, 7, 9, 11, 13] - total: 7
|
|
{TimestampMs: 15_000, Value: 8}, // [15, 17, 19, 21, 23, 25, 27, 29] - total: 8
|
|
{TimestampMs: 30_000, Value: 7}, // [31, 33, 35, 37, 39, 41, 43] - total: 7
|
|
{TimestampMs: 45_000, Value: 3}, // [45, 47, 49] - total: 3
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
},
|
|
expectedL2: []*tempopb.TimeSeries{
|
|
{
|
|
PromLabels: `{".service.name"="even"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueString(".service.name", "even")},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 2 * 7},
|
|
{TimestampMs: 15_000, Value: 2 * 7},
|
|
{TimestampMs: 30_000, Value: 2 * 8},
|
|
{TimestampMs: 45_000, Value: 2 * 2},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
{
|
|
PromLabels: `{".service.name"="odd"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueString(".service.name", "odd")},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 2 * 7},
|
|
{TimestampMs: 15_000, Value: 2 * 8},
|
|
{TimestampMs: 30_000, Value: 2 * 7},
|
|
{TimestampMs: 45_000, Value: 2 * 3},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
},
|
|
},
|
|
{
|
|
name: "min_over_time",
|
|
req: requestWithDefaultRange("{ } | min_over_time(duration)"),
|
|
expectedL1: []*tempopb.TimeSeries{
|
|
{
|
|
PromLabels: `{__name__="min_over_time"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueString("__name__", "min_over_time")},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 1}, // Interval [1, 14], min is 1
|
|
{TimestampMs: 15_000, Value: 15}, // Interval [15, 29], min is 15
|
|
{TimestampMs: 30_000, Value: 30}, // Interval [30, 44], min is 30
|
|
{TimestampMs: 45_000, Value: 45}, // Interval [45, 50), min is 45
|
|
},
|
|
},
|
|
},
|
|
expectedL2: nil, // results should be the same: min(a, a) = a
|
|
},
|
|
{
|
|
name: "max_over_time",
|
|
req: requestWithDefaultRange("{ } | max_over_time(duration)"),
|
|
expectedL1: []*tempopb.TimeSeries{
|
|
{
|
|
PromLabels: `{__name__="max_over_time"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueString("__name__", "max_over_time")},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 14}, // Interval [1, 14], max is 14
|
|
{TimestampMs: 15_000, Value: 29}, // Interval [15, 29], max is 29
|
|
{TimestampMs: 30_000, Value: 44}, // Interval [30, 44], max is 44
|
|
{TimestampMs: 45_000, Value: 49}, // Interval [45, 50), max is 49
|
|
},
|
|
},
|
|
},
|
|
expectedL2: nil, // results should be the same: max(a, a) = a
|
|
},
|
|
{
|
|
name: "avg_over_time",
|
|
req: requestWithDefaultRange("{ } | avg_over_time(duration)"),
|
|
expectedL1: []*tempopb.TimeSeries{
|
|
{
|
|
PromLabels: `{__name__="avg_over_time"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueString("__name__", "avg_over_time")},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 105 / 14.0}, // sum from 1 to 14 is 105
|
|
{TimestampMs: 15_000, Value: 330 / 15.0}, // sum from 15 to 29 is 330
|
|
{TimestampMs: 30_000, Value: 555 / 15.0}, // sum from 30 to 44 is 555
|
|
{TimestampMs: 45_000, Value: 235 / 5.0}, // sum from 45 to 49 is 235
|
|
},
|
|
},
|
|
{
|
|
PromLabels: `{__meta_type="__count", __name__="avg_over_time"}`,
|
|
Labels: []common_v1.KeyValue{
|
|
tempopb.MakeKeyValueString("__name__", "avg_over_time"),
|
|
tempopb.MakeKeyValueString("__meta_type", "__count"),
|
|
},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 14},
|
|
{TimestampMs: 15_000, Value: 15},
|
|
{TimestampMs: 30_000, Value: 15},
|
|
{TimestampMs: 45_000, Value: 5},
|
|
},
|
|
},
|
|
},
|
|
expectedL2: nil,
|
|
expectedL3: []*tempopb.TimeSeries{
|
|
{
|
|
PromLabels: `{__name__="avg_over_time"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueString("__name__", "avg_over_time")},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 105 / 14.0}, // sum from 1 to 14 is 105
|
|
{TimestampMs: 15_000, Value: 330 / 15.0}, // sum from 15 to 29 is 330
|
|
{TimestampMs: 30_000, Value: 555 / 15.0}, // sum from 30 to 44 is 555
|
|
{TimestampMs: 45_000, Value: 235 / 5.0}, // sum from 45 to 49 is 235
|
|
},
|
|
},
|
|
},
|
|
},
|
|
{
|
|
name: "sum_over_time",
|
|
req: requestWithDefaultRange("{ } | sum_over_time(duration)"),
|
|
expectedL1: []*tempopb.TimeSeries{
|
|
{
|
|
PromLabels: `{__name__="sum_over_time"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueString("__name__", "sum_over_time")},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 105}, // sum from 1 to 14 is 105
|
|
{TimestampMs: 15_000, Value: 330}, // sum from 15 to 29 is 330
|
|
{TimestampMs: 30_000, Value: 555}, // sum from 30 to 44 is 555
|
|
{TimestampMs: 45_000, Value: 235}, // sum from 45 to 49 is 235
|
|
},
|
|
},
|
|
},
|
|
expectedL2: []*tempopb.TimeSeries{
|
|
{
|
|
PromLabels: `{__name__="sum_over_time"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueString("__name__", "sum_over_time")},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 2 * 105},
|
|
{TimestampMs: 15_000, Value: 2 * 330},
|
|
{TimestampMs: 30_000, Value: 2 * 555},
|
|
{TimestampMs: 45_000, Value: 2 * 235},
|
|
},
|
|
},
|
|
},
|
|
},
|
|
{
|
|
name: "quantile_over_time",
|
|
req: requestWithDefaultRange("{ } | quantile_over_time(duration, .5)"),
|
|
// first two levels for quantile are buckets and count, then on level 3 we can compute the quantile
|
|
expectedL1: []*tempopb.TimeSeries{
|
|
{
|
|
PromLabels: `{__bucket="1.073741824"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueDouble("__bucket", 1.073741824)},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 1}, // 1 number (1) is less than 1.07
|
|
{TimestampMs: 15_000, Value: 0},
|
|
{TimestampMs: 30_000, Value: 0},
|
|
{TimestampMs: 45_000, Value: 0},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
{
|
|
PromLabels: `{__bucket="2.147483648"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueDouble("__bucket", 2.147483648)},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 1}, // 1 number (2) is between 1.07 and 2.15
|
|
{TimestampMs: 15_000, Value: 0},
|
|
{TimestampMs: 30_000, Value: 0},
|
|
{TimestampMs: 45_000, Value: 0},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
{
|
|
PromLabels: `{__bucket="4.294967296"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueDouble("__bucket", 4.294967296)},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 2}, // 2 numbers (3, 4) are between 2.15 and 4.29
|
|
{TimestampMs: 15_000, Value: 0},
|
|
{TimestampMs: 30_000, Value: 0},
|
|
{TimestampMs: 45_000, Value: 0},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
{
|
|
PromLabels: `{__bucket="8.589934592"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueDouble("__bucket", 8.589934592)},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 4}, // 5, 6, 7, 8
|
|
{TimestampMs: 15_000, Value: 0},
|
|
{TimestampMs: 30_000, Value: 0},
|
|
{TimestampMs: 45_000, Value: 0},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
{
|
|
PromLabels: `{__bucket="17.179869184"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueDouble("__bucket", 17.179869184)},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 6}, // 9, 10, 11, 12, 13, 14
|
|
{TimestampMs: 15_000, Value: 3},
|
|
{TimestampMs: 30_000, Value: 0},
|
|
{TimestampMs: 45_000, Value: 0},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
{
|
|
PromLabels: `{__bucket="34.359738368"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueDouble("__bucket", 34.359738368)},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 0},
|
|
{TimestampMs: 15_000, Value: 12}, // and so on
|
|
{TimestampMs: 30_000, Value: 5},
|
|
{TimestampMs: 45_000, Value: 0},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
{
|
|
PromLabels: `{__bucket="68.719476736"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueDouble("__bucket", 68.719476736)},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 0},
|
|
{TimestampMs: 15_000, Value: 0},
|
|
{TimestampMs: 30_000, Value: 10},
|
|
{TimestampMs: 45_000, Value: 5},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
},
|
|
expectedL2: []*tempopb.TimeSeries{
|
|
{
|
|
PromLabels: `{__bucket="1.073741824"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueDouble("__bucket", 1.073741824)},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 2 * 1},
|
|
{TimestampMs: 15_000, Value: 0},
|
|
{TimestampMs: 30_000, Value: 0},
|
|
{TimestampMs: 45_000, Value: 0},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
{
|
|
PromLabels: `{__bucket="2.147483648"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueDouble("__bucket", 2.147483648)},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 2 * 1},
|
|
{TimestampMs: 15_000, Value: 0},
|
|
{TimestampMs: 30_000, Value: 0},
|
|
{TimestampMs: 45_000, Value: 0},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
{
|
|
PromLabels: `{__bucket="4.294967296"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueDouble("__bucket", 4.294967296)},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 2 * 2},
|
|
{TimestampMs: 15_000, Value: 0},
|
|
{TimestampMs: 30_000, Value: 0},
|
|
{TimestampMs: 45_000, Value: 0},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
{
|
|
PromLabels: `{__bucket="8.589934592"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueDouble("__bucket", 8.589934592)},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 2 * 4},
|
|
{TimestampMs: 15_000, Value: 0},
|
|
{TimestampMs: 30_000, Value: 0},
|
|
{TimestampMs: 45_000, Value: 0},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
{
|
|
PromLabels: `{__bucket="17.179869184"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueDouble("__bucket", 17.179869184)},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 2 * 6},
|
|
{TimestampMs: 15_000, Value: 2 * 3},
|
|
{TimestampMs: 30_000, Value: 0},
|
|
{TimestampMs: 45_000, Value: 0},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
{
|
|
PromLabels: `{__bucket="34.359738368"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueDouble("__bucket", 34.359738368)},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 0},
|
|
{TimestampMs: 15_000, Value: 2 * 12},
|
|
{TimestampMs: 30_000, Value: 2 * 5},
|
|
{TimestampMs: 45_000, Value: 0},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
{
|
|
PromLabels: `{__bucket="68.719476736"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueDouble("__bucket", 68.719476736)},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 0},
|
|
{TimestampMs: 15_000, Value: 0},
|
|
{TimestampMs: 30_000, Value: 2 * 10},
|
|
{TimestampMs: 45_000, Value: 2 * 5},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
},
|
|
expectedL3: []*tempopb.TimeSeries{
|
|
{
|
|
PromLabels: `{p="0.5"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueDouble("p", 0.5)},
|
|
Samples: []tempopb.Sample{
|
|
// Actual q50 is 7.5: half of the number is less than 7.5 and half is greater
|
|
// 1 2 3 4 5 6 7 || 8 9 10 11 12 13 14
|
|
{TimestampMs: 0, Value: 7.22324520567668},
|
|
// 15 16 17 18 19 20 21 || 22 || 23 24 25 26 27 28 29
|
|
{TimestampMs: 15_000, Value: 22.279533901541168},
|
|
// Actual q50 is 37. On low number of samples, the quantile can be not accurate
|
|
// 30 31 32 33 34 35 36 || 37 || 38 39 40 41 42 43 44
|
|
{TimestampMs: 30_000, Value: 40.86084533685759},
|
|
// Actual q50 is 47
|
|
// 45 46 || 47 || 48 49
|
|
{TimestampMs: 45_000, Value: 48.592007999616804},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
},
|
|
},
|
|
{
|
|
name: "histogram_over_time",
|
|
req: requestWithDefaultRange("{ } | histogram_over_time(duration)"),
|
|
expectedL1: []*tempopb.TimeSeries{
|
|
{
|
|
PromLabels: `{__bucket="1.073741824"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueDouble("__bucket", 1.073741824)},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 1},
|
|
{TimestampMs: 15_000, Value: 0},
|
|
{TimestampMs: 30_000, Value: 0},
|
|
{TimestampMs: 45_000, Value: 0},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
{
|
|
PromLabels: `{__bucket="2.147483648"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueDouble("__bucket", 2.147483648)},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 1},
|
|
{TimestampMs: 15_000, Value: 0},
|
|
{TimestampMs: 30_000, Value: 0},
|
|
{TimestampMs: 45_000, Value: 0},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
{
|
|
PromLabels: `{__bucket="4.294967296"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueDouble("__bucket", 4.294967296)},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 2},
|
|
{TimestampMs: 15_000, Value: 0},
|
|
{TimestampMs: 30_000, Value: 0},
|
|
{TimestampMs: 45_000, Value: 0},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
{
|
|
PromLabels: `{__bucket="8.589934592"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueDouble("__bucket", 8.589934592)},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 4},
|
|
{TimestampMs: 15_000, Value: 0},
|
|
{TimestampMs: 30_000, Value: 0},
|
|
{TimestampMs: 45_000, Value: 0},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
{
|
|
PromLabels: `{__bucket="17.179869184"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueDouble("__bucket", 17.179869184)},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 6},
|
|
{TimestampMs: 15_000, Value: 3},
|
|
{TimestampMs: 30_000, Value: 0},
|
|
{TimestampMs: 45_000, Value: 0},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
{
|
|
PromLabels: `{__bucket="34.359738368"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueDouble("__bucket", 34.359738368)},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 0},
|
|
{TimestampMs: 15_000, Value: 12},
|
|
{TimestampMs: 30_000, Value: 5},
|
|
{TimestampMs: 45_000, Value: 0},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
{
|
|
PromLabels: `{__bucket="68.719476736"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueDouble("__bucket", 68.719476736)},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 0},
|
|
{TimestampMs: 15_000, Value: 0},
|
|
{TimestampMs: 30_000, Value: 10},
|
|
{TimestampMs: 45_000, Value: 5},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
},
|
|
expectedL2: []*tempopb.TimeSeries{
|
|
{
|
|
PromLabels: `{__bucket="1.073741824"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueDouble("__bucket", 1.073741824)},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 2 * 1},
|
|
{TimestampMs: 15_000, Value: 0},
|
|
{TimestampMs: 30_000, Value: 0},
|
|
{TimestampMs: 45_000, Value: 0},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
{
|
|
PromLabels: `{__bucket="2.147483648"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueDouble("__bucket", 2.147483648)},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 2 * 1},
|
|
{TimestampMs: 15_000, Value: 0},
|
|
{TimestampMs: 30_000, Value: 0},
|
|
{TimestampMs: 45_000, Value: 0},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
{
|
|
PromLabels: `{__bucket="4.294967296"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueDouble("__bucket", 4.294967296)},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 2 * 2},
|
|
{TimestampMs: 15_000, Value: 0},
|
|
{TimestampMs: 30_000, Value: 0},
|
|
{TimestampMs: 45_000, Value: 0},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
{
|
|
PromLabels: `{__bucket="8.589934592"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueDouble("__bucket", 8.589934592)},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 2 * 4},
|
|
{TimestampMs: 15_000, Value: 0},
|
|
{TimestampMs: 30_000, Value: 0},
|
|
{TimestampMs: 45_000, Value: 0},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
{
|
|
PromLabels: `{__bucket="17.179869184"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueDouble("__bucket", 17.179869184)},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 2 * 6},
|
|
{TimestampMs: 15_000, Value: 2 * 3},
|
|
{TimestampMs: 30_000, Value: 0},
|
|
{TimestampMs: 45_000, Value: 0},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
{
|
|
PromLabels: `{__bucket="34.359738368"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueDouble("__bucket", 34.359738368)},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 0},
|
|
{TimestampMs: 15_000, Value: 2 * 12},
|
|
{TimestampMs: 30_000, Value: 2 * 5},
|
|
{TimestampMs: 45_000, Value: 0},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
{
|
|
PromLabels: `{__bucket="68.719476736"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueDouble("__bucket", 68.719476736)},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 0},
|
|
{TimestampMs: 15_000, Value: 0},
|
|
{TimestampMs: 30_000, Value: 2 * 10},
|
|
{TimestampMs: 45_000, Value: 2 * 5},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
},
|
|
},
|
|
{
|
|
name: "compare_no_spans",
|
|
req: requestWithDefaultRange(`{ .service.name="does_not_exist" } | compare({ })`),
|
|
expectedL1: []*tempopb.TimeSeries{},
|
|
},
|
|
{
|
|
name: "compare_no_spans",
|
|
req: requestWithDefaultRange(`{ .service.name="does_not_exist" } | compare({ .service.name="does_not_exist_for_sure" })`),
|
|
expectedL1: []*tempopb.TimeSeries{},
|
|
},
|
|
// --- Non-standard range queries ---
|
|
{
|
|
name: "end<step",
|
|
req: &tempopb.QueryRangeRequest{
|
|
Start: 1,
|
|
End: 3 * uint64(time.Second),
|
|
Step: 5 * uint64(time.Second),
|
|
Query: `{ } | count_over_time()`,
|
|
},
|
|
expectedL1: []*tempopb.TimeSeries{
|
|
{
|
|
PromLabels: `{__name__="count_over_time"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueString("__name__", "count_over_time")},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 2}, // 1, 2
|
|
{TimestampMs: 5_000, Value: 0},
|
|
},
|
|
},
|
|
},
|
|
expectedL2: []*tempopb.TimeSeries{
|
|
{
|
|
PromLabels: `{__name__="count_over_time"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueString("__name__", "count_over_time")},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 2 * 2},
|
|
{TimestampMs: 5_000, Value: 0},
|
|
},
|
|
},
|
|
},
|
|
},
|
|
{
|
|
name: "end=step",
|
|
req: &tempopb.QueryRangeRequest{
|
|
Start: 1,
|
|
End: 5 * uint64(time.Second),
|
|
Step: 5 * uint64(time.Second),
|
|
Query: `{ } | count_over_time()`,
|
|
},
|
|
expectedL1: []*tempopb.TimeSeries{
|
|
{
|
|
PromLabels: `{__name__="count_over_time"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueString("__name__", "count_over_time")},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 4}, // 1, 2, 3, 4
|
|
{TimestampMs: 5_000, Value: 0},
|
|
},
|
|
},
|
|
},
|
|
expectedL2: []*tempopb.TimeSeries{
|
|
{
|
|
PromLabels: `{__name__="count_over_time"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueString("__name__", "count_over_time")},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 2 * 4},
|
|
{TimestampMs: 5_000, Value: 0},
|
|
},
|
|
},
|
|
},
|
|
},
|
|
{
|
|
name: "small step",
|
|
req: &tempopb.QueryRangeRequest{
|
|
Start: 1,
|
|
End: 3 * uint64(time.Second),
|
|
Step: 500 * uint64(time.Millisecond),
|
|
Query: `{ } | count_over_time()`,
|
|
},
|
|
expectedL1: []*tempopb.TimeSeries{
|
|
{
|
|
PromLabels: `{__name__="count_over_time"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueString("__name__", "count_over_time")},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 0},
|
|
{TimestampMs: 500, Value: 0},
|
|
{TimestampMs: 1000, Value: 1},
|
|
{TimestampMs: 1500, Value: 0},
|
|
{TimestampMs: 2000, Value: 1},
|
|
{TimestampMs: 2500, Value: 0},
|
|
{TimestampMs: 3000, Value: 0},
|
|
},
|
|
},
|
|
},
|
|
expectedL2: []*tempopb.TimeSeries{
|
|
{
|
|
PromLabels: `{__name__="count_over_time"}`,
|
|
Labels: []common_v1.KeyValue{tempopb.MakeKeyValueString("__name__", "count_over_time")},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 0},
|
|
{TimestampMs: 500, Value: 0},
|
|
{TimestampMs: 1000, Value: 2 * 1},
|
|
{TimestampMs: 1500, Value: 0},
|
|
{TimestampMs: 2000, Value: 2 * 1},
|
|
{TimestampMs: 2500, Value: 0},
|
|
{TimestampMs: 3000, Value: 0},
|
|
},
|
|
},
|
|
},
|
|
},
|
|
}
|
|
|
|
var expectedCompareTs = []*tempopb.TimeSeries{
|
|
{
|
|
PromLabels: `{__meta_type="baseline", "resource.service.name"="odd"}`,
|
|
Labels: []common_v1.KeyValue{
|
|
tempopb.MakeKeyValueString("__meta_type", "baseline"),
|
|
tempopb.MakeKeyValueString("resource.service.name", "odd"),
|
|
},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 7},
|
|
{TimestampMs: 15_000, Value: 8},
|
|
{TimestampMs: 30_000, Value: 7},
|
|
{TimestampMs: 45_000, Value: 3},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
{
|
|
PromLabels: `{__meta_type="baseline_total", "resource.service.name"="<nil>"}`,
|
|
Labels: []common_v1.KeyValue{
|
|
tempopb.MakeKeyValueString("__meta_type", "baseline_total"),
|
|
tempopb.MakeKeyValueString("resource.service.name", "nil"),
|
|
},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 7},
|
|
{TimestampMs: 15_000, Value: 8},
|
|
{TimestampMs: 30_000, Value: 7},
|
|
{TimestampMs: 45_000, Value: 3},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
{
|
|
PromLabels: `{__meta_type="selection", "resource.service.name"="even"}`,
|
|
Labels: []common_v1.KeyValue{
|
|
tempopb.MakeKeyValueString("__meta_type", "selection"),
|
|
tempopb.MakeKeyValueString("resource.service.name", "even"),
|
|
},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 7},
|
|
{TimestampMs: 15_000, Value: 7},
|
|
{TimestampMs: 30_000, Value: 8},
|
|
{TimestampMs: 45_000, Value: 2},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
{
|
|
PromLabels: `{__meta_type="selection_total", "resource.service.name"="<nil>"}`,
|
|
Labels: []common_v1.KeyValue{
|
|
tempopb.MakeKeyValueString("__meta_type", "selection_total"),
|
|
tempopb.MakeKeyValueString("resource.service.name", "nil"),
|
|
},
|
|
Samples: []tempopb.Sample{
|
|
{TimestampMs: 0, Value: 7},
|
|
{TimestampMs: 15_000, Value: 7},
|
|
{TimestampMs: 30_000, Value: 8},
|
|
{TimestampMs: 45_000, Value: 2},
|
|
{TimestampMs: 60_000, Value: 0},
|
|
},
|
|
},
|
|
}
|
|
|
|
// TestTempoDBQueryRange tests the metrics query functionality of TempoDB by verifying various types of
|
|
// time-series queries on trace data. The test:
|
|
//
|
|
// 1. Sets up a test environment
|
|
//
|
|
// 2. Generates test data:
|
|
// - 100 test spans distributed across 1-100 seconds
|
|
// - Each span's duration equals its start time (e.g. span at 2s has duration 2s)
|
|
// - Spans are tagged with service names "even" or "odd" based on their start time
|
|
//
|
|
// 3. Tests various query types:
|
|
// - Rate queries (rate(), rate() with filters)
|
|
// - Count queries (count_over_time(), count_over_time() by service)
|
|
// - Statistical queries (min, max, avg, sum, quantile, histogram)
|
|
// - Edge cases with different time ranges and step sizes
|
|
//
|
|
// 4. Validates results at three processing levels covering the whole query pipeline:
|
|
// - Level 1: Initial query results
|
|
// - Level 2: Results after first aggregation simulating multiple sources
|
|
// - Level 3: Final aggregation results
|
|
func TestTempoDBQueryRange(t *testing.T) {
|
|
var (
|
|
tempDir = t.TempDir()
|
|
blockVersion = vparquet4.VersionString
|
|
)
|
|
|
|
dc := backend.DedicatedColumns{
|
|
{Scope: "resource", Name: "res-dedicated.01", Type: "string"},
|
|
{Scope: "resource", Name: "res-dedicated.02", Type: "string"},
|
|
{Scope: "span", Name: "span-dedicated.01", Type: "string"},
|
|
{Scope: "span", Name: "span-dedicated.02", Type: "string"},
|
|
}
|
|
r, w, c, err := New(&Config{
|
|
Backend: backend.Local,
|
|
Local: &local.Config{
|
|
Path: path.Join(tempDir, "traces"),
|
|
},
|
|
Block: &common.BlockConfig{
|
|
IndexDownsampleBytes: 17,
|
|
BloomFP: .01,
|
|
BloomShardSizeBytes: 100_000,
|
|
Version: blockVersion,
|
|
IndexPageSizeBytes: 1000,
|
|
RowGroupSizeBytes: 10000,
|
|
DedicatedColumns: dc,
|
|
},
|
|
WAL: &wal.Config{
|
|
Filepath: path.Join(tempDir, "wal"),
|
|
IngestionSlack: time.Since(time.Time{}),
|
|
},
|
|
Search: &SearchConfig{
|
|
ChunkSizeBytes: 1_000_000,
|
|
ReadBufferCount: 8, ReadBufferSizeBytes: 4 * 1024 * 1024,
|
|
},
|
|
BlocklistPoll: 0,
|
|
}, nil, log.NewNopLogger())
|
|
require.NoError(t, err)
|
|
|
|
err = c.EnableCompaction(context.Background(), &CompactorConfig{
|
|
ChunkSizeBytes: 10,
|
|
MaxCompactionRange: time.Hour,
|
|
BlockRetention: 0,
|
|
CompactedBlockRetention: 0,
|
|
}, &mockSharder{}, &mockOverrides{})
|
|
require.NoError(t, err)
|
|
|
|
ctx := context.Background()
|
|
r.EnablePolling(ctx, &mockJobSharder{})
|
|
|
|
// Write to wal
|
|
wal := w.WAL()
|
|
|
|
meta := &backend.BlockMeta{BlockID: backend.NewUUID(), TenantID: testTenantID, DedicatedColumns: dc}
|
|
head, err := wal.NewBlock(meta, model.CurrentEncoding)
|
|
require.NoError(t, err)
|
|
dec := model.MustNewSegmentDecoder(model.CurrentEncoding)
|
|
|
|
totalSpans := 100
|
|
for i := 1; i <= totalSpans; i++ {
|
|
tid := test.ValidTraceID(nil)
|
|
|
|
sp := test.MakeSpan(tid)
|
|
|
|
// Start time is i seconds
|
|
sp.StartTimeUnixNano = uint64(i * int(time.Second))
|
|
|
|
// Duration is i seconds
|
|
sp.EndTimeUnixNano = sp.StartTimeUnixNano + uint64(i*int(time.Second))
|
|
|
|
// Service name
|
|
var svcName string
|
|
if i%2 == 0 {
|
|
svcName = "even"
|
|
} else {
|
|
svcName = "odd"
|
|
}
|
|
|
|
tr := &tempopb.Trace{
|
|
ResourceSpans: []*v1.ResourceSpans{
|
|
{
|
|
Resource: &resource_v1.Resource{
|
|
Attributes: []*common_v1.KeyValue{tempopb.MakeKeyValueStringPtr("service.name", svcName)},
|
|
},
|
|
ScopeSpans: []*v1.ScopeSpans{
|
|
{
|
|
Spans: []*v1.Span{
|
|
sp,
|
|
},
|
|
},
|
|
},
|
|
},
|
|
},
|
|
}
|
|
|
|
b1, err := dec.PrepareForWrite(tr, 0, 0)
|
|
require.NoError(t, err)
|
|
|
|
b2, err := dec.ToObject([][]byte{b1})
|
|
require.NoError(t, err)
|
|
err = head.Append(tid, b2, 0, 0, true)
|
|
require.NoError(t, err)
|
|
}
|
|
|
|
// Complete block
|
|
block, err := w.CompleteBlock(context.Background(), head)
|
|
require.NoError(t, err)
|
|
|
|
f := traceql.NewSpansetFetcherWrapper(func(ctx context.Context, req traceql.FetchSpansRequest) (traceql.FetchSpansResponse, error) {
|
|
return block.Fetch(ctx, req, common.DefaultSearchOptions())
|
|
})
|
|
|
|
for _, tc := range queryRangeTestCases {
|
|
t.Run(tc.name, func(t *testing.T) {
|
|
e := traceql.NewEngine()
|
|
eval, err := e.CompileMetricsQueryRange(tc.req, 0, 0, false)
|
|
require.NoError(t, err)
|
|
|
|
err = eval.Do(ctx, f, 0, 0, 0)
|
|
require.NoError(t, err)
|
|
|
|
actual := eval.Results().ToProto(tc.req)
|
|
expected := tc.expectedL1
|
|
|
|
// Slice order is not deterministic, so we sort the slices before comparing
|
|
sortTimeSeries(actual)
|
|
sortTimeSeries(expected)
|
|
|
|
if diff := cmp.Diff(expected, actual, floatComparer); diff != "" {
|
|
t.Errorf("Unexpected results for Level 1 processing. Query: %v\n Diff: %v", tc.req.Query, diff)
|
|
}
|
|
|
|
evalLevel2, err := e.CompileMetricsQueryRangeNonRaw(tc.req, traceql.AggregateModeSum)
|
|
require.NoError(t, err)
|
|
evalLevel2.ObserveSeries(actual)
|
|
evalLevel2.ObserveSeries(actual) // emulate merging from two sources
|
|
actual = evalLevel2.Results().ToProto(tc.req)
|
|
sortTimeSeries(actual)
|
|
|
|
if tc.expectedL2 != nil {
|
|
expected = tc.expectedL2
|
|
sortTimeSeries(expected)
|
|
}
|
|
|
|
if diff := cmp.Diff(expected, actual, floatComparer); diff != "" {
|
|
t.Errorf("Unexpected results for Level 2 processing. Query: %v\n Diff: %v", tc.req.Query, diff)
|
|
}
|
|
|
|
evalLevel3, err := e.CompileMetricsQueryRangeNonRaw(tc.req, traceql.AggregateModeFinal)
|
|
require.NoError(t, err)
|
|
evalLevel3.ObserveSeries(actual)
|
|
actual = evalLevel3.Results().ToProto(tc.req)
|
|
sortTimeSeries(actual)
|
|
|
|
if tc.expectedL3 != nil {
|
|
expected = tc.expectedL3
|
|
sortTimeSeries(expected)
|
|
}
|
|
|
|
if diff := cmp.Diff(expected, actual, floatComparer); diff != "" {
|
|
t.Errorf("Unexpected results for Level 3 processing. Query: %v\n Diff: %v", tc.req.Query, diff)
|
|
}
|
|
})
|
|
}
|
|
|
|
t.Run("compare", func(t *testing.T) {
|
|
// compare operation generates enormous amount of time series,
|
|
// so we filter by service.name to test at least part of the results
|
|
req := requestWithDefaultRange(`{} | compare({ .service.name="even" })`)
|
|
e := traceql.NewEngine()
|
|
|
|
// Level 1
|
|
eval, err := e.CompileMetricsQueryRange(req, 0, 0, false)
|
|
require.NoError(t, err)
|
|
|
|
err = eval.Do(ctx, f, 0, 0, 0)
|
|
require.NoError(t, err)
|
|
|
|
actual := eval.Results().ToProto(req)
|
|
|
|
const pattern = `"resource.service.name"=`
|
|
targetTs := filterTimeSeriesByPromLabel(actual, pattern)
|
|
sortTimeSeries(targetTs)
|
|
require.Equal(t, expectedCompareTs, targetTs)
|
|
|
|
// Level 2
|
|
evalLevel2, err := e.CompileMetricsQueryRangeNonRaw(req, traceql.AggregateModeSum)
|
|
require.NoError(t, err)
|
|
evalLevel2.ObserveSeries(actual)
|
|
actual = evalLevel2.Results().ToProto(req)
|
|
|
|
targetTs = filterTimeSeriesByPromLabel(actual, pattern)
|
|
sortTimeSeries(targetTs)
|
|
require.Equal(t, expectedCompareTs, targetTs)
|
|
|
|
// Level 3
|
|
evalLevel3, err := e.CompileMetricsQueryRangeNonRaw(req, traceql.AggregateModeFinal)
|
|
require.NoError(t, err)
|
|
evalLevel3.ObserveSeries(actual)
|
|
actual = evalLevel3.Results().ToProto(req)
|
|
|
|
targetTs = filterTimeSeriesByPromLabel(actual, pattern)
|
|
sortTimeSeries(targetTs)
|
|
require.Equal(t, expectedCompareTs, targetTs)
|
|
})
|
|
}
|
|
|
|
func sortTimeSeries(ts []*tempopb.TimeSeries) {
|
|
sort.Slice(ts, func(i, j int) bool {
|
|
return ts[i].PromLabels < ts[j].PromLabels
|
|
})
|
|
}
|
|
|
|
var floatComparer = cmp.Comparer(func(x, y float64) bool {
|
|
return math.Abs(x-y) < 1e-6
|
|
})
|
|
|
|
// filterTimeSeries filters the time series by the pattern in the PromLabels
|
|
func filterTimeSeriesByPromLabel(ts []*tempopb.TimeSeries, pattern string) []*tempopb.TimeSeries {
|
|
var targetTs []*tempopb.TimeSeries
|
|
for _, ts := range ts {
|
|
if strings.Contains(ts.PromLabels, pattern) {
|
|
targetTs = append(targetTs, ts)
|
|
}
|
|
}
|
|
return targetTs
|
|
}
|