Infer max query downsample resolution from promql query (#7012)

* Adjust max_source_resolution automatically based promql queries

Signed-off-by: Ben Ye <[email protected]>

* fix data race

Signed-off-by: yeya24 <[email protected]>

---------

Signed-off-by: Ben Ye <[email protected]>
Signed-off-by: yeya24 <[email protected]>

Author: yeya24

CHANGELOG.md

@@ -13,7 +13,6 @@ We use *breaking :warning:* to mark changes that are not backward compatible (re
 ### Fixed
 - [#8091](https://github.com/thanos-io/thanos/pull/8091) *: Add POST into allowed CORS methods header
 - [#8046](https://github.com/thanos-io/thanos/pull/8046) Query-Frontend: Fix query statistic reporting for range queries when caching is enabled.
-
 - [#7978](https://github.com/thanos-io/thanos/pull/7978) Receive: Fix deadlock during local writes when `split-tenant-label-name` is used
 - [#8016](https://github.com/thanos-io/thanos/pull/8016) Query Frontend: Fix @ modifier not being applied correctly on sub queries.
 
@@ -28,6 +27,7 @@ We use *breaking :warning:* to mark changes that are not backward compatible (re
 ### Changed
 
 - [#7890](https://github.com/thanos-io/thanos/pull/7890) Query,Ruler: *breaking :warning:* deprecated `--store.sd-file` and `--store.sd-interval` to be replaced with `--endpoint.sd-config` and `--endpoint-sd-config-reload-interval`; removed legacy flags to pass endpoints `--store`, `--metadata`, `--rule`, `--exemplar`.
+- [#7012](https://github.com/thanos-io/thanos/pull/7012) Query: Automatically adjust `max_source_resolution` based on promql query to avoid querying data from higher resolution resulting empty results.
 
 ### Removed
 

pkg/query/querier.go

@@ -27,6 +27,18 @@ import (
 	"github.com/thanos-io/thanos/pkg/tracing"
 )
 
+var promqlFuncRequiresTwoSamples = map[string]struct{}{
+	"rate":                         {},
+	"irate":                        {},
+	"increase":                     {},
+	"delta":                        {},
+	"idelta":                       {},
+	"deriv":                        {},
+	"predict_linear":               {},
+	"holt_winters":                 {},
+	"double_exponential_smoothing": {},
+}
+
 type seriesStatsReporter func(seriesStats storepb.SeriesStatsCounter)
 
 var NoopSeriesStatsReporter seriesStatsReporter = func(_ storepb.SeriesStatsCounter) {}
@@ -320,6 +332,7 @@ func (q *querier) selectFn(ctx context.Context, hints *storage.SelectHints, ms .
 	}
 
 	aggrs := aggrsFromFunc(hints.Func)
+	maxResolutionMillis := maxResolutionFromSelectHints(q.maxResolutionMillis, hints.Range, hints.Func)
 
 	// TODO(bwplotka): Pass it using the SeriesRequest instead of relying on context.
 	ctx = context.WithValue(ctx, store.StoreMatcherKey, q.storeDebugMatchers)
@@ -333,7 +346,7 @@ func (q *querier) selectFn(ctx context.Context, hints *storage.SelectHints, ms .
 		MaxTime:                 hints.End,
 		Limit:                   int64(hints.Limit),
 		Matchers:                sms,
-		MaxResolutionWindow:     q.maxResolutionMillis,
+		MaxResolutionWindow:     maxResolutionMillis,
 		Aggregates:              aggrs,
 		ShardInfo:               q.shardInfo,
 		PartialResponseStrategy: q.partialResponseStrategy,
@@ -460,3 +473,13 @@ func (q *querier) LabelNames(ctx context.Context, hints *storage.LabelHints, mat
 }
 
 func (q *querier) Close() error { return nil }
+
+// maxResolutionFromSelectHints finds the max possible resolution by inferring from the promql query.
+func maxResolutionFromSelectHints(maxResolutionMillis int64, hintsRange int64, hintsFunc string) int64 {
+	if hintsRange > 0 {
+		if _, ok := promqlFuncRequiresTwoSamples[hintsFunc]; ok {
+			maxResolutionMillis = min(maxResolutionMillis, hintsRange/2)
+		}
+	}
+	return maxResolutionMillis
+}

pkg/query/querier_test.go

@@ -32,6 +32,7 @@ import (
 	"github.com/prometheus/prometheus/util/gate"
 	"github.com/thanos-io/thanos/pkg/logutil"
 
+	"github.com/thanos-io/thanos/pkg/compact/downsample"
 	"github.com/thanos-io/thanos/pkg/component"
 	"github.com/thanos-io/thanos/pkg/store"
 	"github.com/thanos-io/thanos/pkg/store/labelpb"
@@ -1279,3 +1280,134 @@ func storeSeriesResponse(t testing.TB, lset labels.Labels, smplChunks ...[]sampl
 	}
 	return storepb.NewSeriesResponse(&s)
 }
+
+func TestMaxResolutionFromSelectHints(t *testing.T) {
+	twoMinRange := int64(2 * 60 * 1000)
+	for _, tc := range []struct {
+		name                string
+		maxResolutionMillis int64
+		hints               storage.SelectHints
+		expected            int64
+	}{
+		{
+			name: "no range",
+			hints: storage.SelectHints{
+				Range: 0,
+			},
+			maxResolutionMillis: downsample.ResLevel1,
+			expected:            downsample.ResLevel1,
+		},
+		{
+			name: "no function",
+			hints: storage.SelectHints{
+				Range: twoMinRange,
+			},
+			maxResolutionMillis: downsample.ResLevel1,
+			expected:            downsample.ResLevel1,
+		},
+		{
+			name: "function doesn't impact resolution",
+			hints: storage.SelectHints{
+				Range: twoMinRange,
+				Func:  "max_over_time",
+			},
+			maxResolutionMillis: downsample.ResLevel1,
+			expected:            downsample.ResLevel1,
+		},
+		{
+			name: "rate",
+			hints: storage.SelectHints{
+				Range: twoMinRange,
+				Func:  "rate",
+			},
+			maxResolutionMillis: downsample.ResLevel1,
+			expected:            twoMinRange / 2,
+		},
+		{
+			name: "rate",
+			hints: storage.SelectHints{
+				Range: twoMinRange,
+				Func:  "rate",
+			},
+			maxResolutionMillis: downsample.ResLevel1,
+			expected:            twoMinRange / 2,
+		},
+		{
+			name: "irate",
+			hints: storage.SelectHints{
+				Range: twoMinRange,
+				Func:  "irate",
+			},
+			maxResolutionMillis: downsample.ResLevel1,
+			expected:            twoMinRange / 2,
+		},
+		{
+			name: "increase",
+			hints: storage.SelectHints{
+				Range: twoMinRange,
+				Func:  "increase",
+			},
+			maxResolutionMillis: downsample.ResLevel1,
+			expected:            twoMinRange / 2,
+		},
+		{
+			name: "delta",
+			hints: storage.SelectHints{
+				Range: twoMinRange,
+				Func:  "delta",
+			},
+			maxResolutionMillis: downsample.ResLevel1,
+			expected:            twoMinRange / 2,
+		},
+		{
+			name: "idelta",
+			hints: storage.SelectHints{
+				Range: twoMinRange,
+				Func:  "idelta",
+			},
+			maxResolutionMillis: downsample.ResLevel1,
+			expected:            twoMinRange / 2,
+		},
+		{
+			name: "deriv",
+			hints: storage.SelectHints{
+				Range: twoMinRange,
+				Func:  "deriv",
+			},
+			maxResolutionMillis: downsample.ResLevel1,
+			expected:            twoMinRange / 2,
+		},
+		{
+			name: "predict_linear",
+			hints: storage.SelectHints{
+				Range: twoMinRange,
+				Func:  "predict_linear",
+			},
+			maxResolutionMillis: downsample.ResLevel1,
+			expected:            twoMinRange / 2,
+		},
+		{
+			name: "holt_winters",
+			hints: storage.SelectHints{
+				Range: twoMinRange,
+				Func:  "holt_winters",
+			},
+			maxResolutionMillis: downsample.ResLevel1,
+			expected:            twoMinRange / 2,
+		},
+		{
+			name: "double_exponential_smoothing",
+			hints: storage.SelectHints{
+				Range: twoMinRange,
+				Func:  "double_exponential_smoothing",
+			},
+			maxResolutionMillis: downsample.ResLevel1,
+			expected:            twoMinRange / 2,
+		},
+	} {
+		t.Run(tc.name, func(t *testing.T) {
+			res := maxResolutionFromSelectHints(tc.maxResolutionMillis, tc.hints.Range, tc.hints.Func)
+			testutil.Equals(t, tc.expected, res)
+		})
+	}
+}