app/vmselect/promql: add share_le_over_time and share_gt_over_time functions for SLI and SLO calculations

app/vmselect/promql/exec_test.go

@@ -3083,6 +3083,28 @@ func TestExecSuccess(t *testing.T) {
 		resultExpected := []netstorage.Result{r}
 		f(q, resultExpected)
 	})
+	t.Run(`share_gt_over_time`, func(t *testing.T) {
+		t.Parallel()
+		q := `share_gt_over_time(rand(0)[200s:10s], 0.7)`
+		r := netstorage.Result{
+			MetricName: metricNameExpected,
+			Values:     []float64{0.35, 0.3, 0.5, 0.3, 0.3, 0.25},
+			Timestamps: timestampsExpected,
+		}
+		resultExpected := []netstorage.Result{r}
+		f(q, resultExpected)
+	})
+	t.Run(`share_le_over_time`, func(t *testing.T) {
+		t.Parallel()
+		q := `share_le_over_time(rand(0)[200s:10s], 0.7)`
+		r := netstorage.Result{
+			MetricName: metricNameExpected,
+			Values:     []float64{0.65, 0.7, 0.5, 0.7, 0.7, 0.75},
+			Timestamps: timestampsExpected,
+		}
+		resultExpected := []netstorage.Result{r}
+		f(q, resultExpected)
+	})
 	t.Run(`increases_over_time`, func(t *testing.T) {
 		t.Parallel()
 		q := `increases_over_time(rand(0)[200s:10s])`

app/vmselect/promql/rollup.go

@@ -49,6 +49,8 @@ var rollupFuncs = map[string]newRollupFunc{
 	"lifetime":            newRollupFuncOneArg(rollupLifetime),
 	"lag":                 newRollupFuncOneArg(rollupLag),
 	"scrape_interval":     newRollupFuncOneArg(rollupScrapeInterval),
+	"share_le_over_time":  newRollupShareLE,
+	"share_gt_over_time":  newRollupShareGT,
 	"rollup":              newRollupFuncOneArg(rollupFake),
 	"rollup_rate":         newRollupFuncOneArg(rollupFake), // + rollupFuncsRemoveCounterResets
 	"rollup_deriv":        newRollupFuncOneArg(rollupFake),
@@ -529,6 +531,56 @@ func linearRegression(rfa *rollupFuncArg) (float64, float64) {
 	return v, k
 }
 
+func newRollupShareLE(args []interface{}) (rollupFunc, error) {
+	return newRollupShareFilter(args, countFilterLE)
+}
+
+func countFilterLE(values []float64, le float64) int {
+	n := 0
+	for _, v := range values {
+		if v <= le {
+			n++
+		}
+	}
+	return n
+}
+
+func newRollupShareGT(args []interface{}) (rollupFunc, error) {
+	return newRollupShareFilter(args, countFilterGT)
+}
+
+func countFilterGT(values []float64, gt float64) int {
+	n := 0
+	for _, v := range values {
+		if v > gt {
+			n++
+		}
+	}
+	return n
+}
+
+func newRollupShareFilter(args []interface{}, countFilter func(values []float64, limit float64) int) (rollupFunc, error) {
+	if err := expectRollupArgsNum(args, 2); err != nil {
+		return nil, err
+	}
+	limits, err := getScalar(args[1], 1)
+	if err != nil {
+		return nil, err
+	}
+	rf := func(rfa *rollupFuncArg) float64 {
+		// There is no need in handling NaNs here, since they must be cleaned up
+		// before calling rollup funcs.
+		values := rfa.values
+		if len(values) == 0 {
+			return nan
+		}
+		limit := limits[rfa.idx]
+		n := countFilter(values, limit)
+		return float64(n) / float64(len(values))
+	}
+	return rf, nil
+}
+
 func newRollupQuantile(args []interface{}) (rollupFunc, error) {
 	if err := expectRollupArgsNum(args, 2); err != nil {
 		return nil, err

app/vmselect/promql/rollup_test.go

@@ -192,6 +192,52 @@ func testRollupFunc(t *testing.T, funcName string, args []interface{}, meExpecte
 	}
 }
 
+func TestRollupShareLEOverTime(t *testing.T) {
+	f := func(le, vExpected float64) {
+		t.Helper()
+		les := []*timeseries{{
+			Values:     []float64{le},
+			Timestamps: []int64{123},
+		}}
+		var me metricsql.MetricExpr
+		args := []interface{}{&metricsql.RollupExpr{Expr: &me}, les}
+		testRollupFunc(t, "share_le_over_time", args, &me, vExpected)
+	}
+
+	f(-123, 0)
+	f(0, 0)
+	f(10, 0)
+	f(12, 0.08333333333333333)
+	f(30, 0.16666666666666666)
+	f(50, 0.75)
+	f(100, 0.9166666666666666)
+	f(123, 1)
+	f(1000, 1)
+}
+
+func TestRollupShareGTOverTime(t *testing.T) {
+	f := func(gt, vExpected float64) {
+		t.Helper()
+		gts := []*timeseries{{
+			Values:     []float64{gt},
+			Timestamps: []int64{123},
+		}}
+		var me metricsql.MetricExpr
+		args := []interface{}{&metricsql.RollupExpr{Expr: &me}, gts}
+		testRollupFunc(t, "share_gt_over_time", args, &me, vExpected)
+	}
+
+	f(-123, 1)
+	f(0, 1)
+	f(10, 1)
+	f(12, 0.9166666666666666)
+	f(30, 0.8333333333333334)
+	f(50, 0.25)
+	f(100, 0.08333333333333333)
+	f(123, 0)
+	f(1000, 0)
+}
+
 func TestRollupQuantileOverTime(t *testing.T) {
 	f := func(phi, vExpected float64) {
 		t.Helper()

docs/ExtendedPromQL.md

@@ -89,3 +89,7 @@ This functionality can be tried at [an editable Grafana dashboard](http://play-g
    - `bottomk_max(k, q)` - returns bottom K time series with the min maximums on the given time range
    - `bottomk_avg(k, q)` - returns bottom K time series with the min averages on the given time range
    - `bottomk_median(k, q)` - returns bottom K time series with the min medians on the given time range
+- `share_le_over_time(m[d], le)` - returns share (in the range 0..1) of values in `m` over `d`, which are smaller or equal to `le`. Useful for calculating SLI and SLO.
+  Example: `share_le_over_time(memory_usage_bytes[24h], 100*1024*1024)` returns the share of time series values for the last 24 hours when memory usage was below or equal to 100MB.
+- `share_gt_over_time(m[d], gt)` - returns share (in the range 0..1) of values in `m` over `d`, which are bigger than `gt`. Useful for calculating SLI and SLO.
+  Example: `share_gt_over_time(up[24h], 0)` - returns service availability for the last 24 hours.

lib/metricsql/rollup.go

@@ -41,6 +41,8 @@ var rollupFuncs = map[string]bool{
 	"lifetime":            true,
 	"lag":                 true,
 	"scrape_interval":     true,
+	"share_le_over_time":  true,
+	"share_gt_over_time":  true,
 	"rollup":              true,
 	"rollup_rate":         true,
 	"rollup_deriv":        true,