metricsql: properly evaluate timezone_offset over time interval (#2842)

* metricsql: properly evaluate `timezone_offset` over time interval

https://github.com/VictoriaMetrics/VictoriaMetrics/issues/2771
Signed-off-by: hagen1778 <roman@victoriametrics.com>

* Update docs/CHANGELOG.md

Co-authored-by: Aliaksandr Valialkin <valyala@victoriametrics.com>

app/vmselect/promql/exec_test.go

@@ -224,10 +224,12 @@ func TestExecSuccess(t *testing.T) {
 	t.Run("timezone_offset(America/New_York)", func(t *testing.T) {
 		t.Parallel()
 		q := `timezone_offset("America/New_York")`
-		offset, err := getTimezoneOffset("America/New_York")
+		loc, err := time.LoadLocation("America/New_York")
 		if err != nil {
 			t.Fatalf("cannot obtain timezone: %s", err)
 		}
+		at := time.Unix(timestampsExpected[0]/1000, 0)
+		_, offset := at.In(loc).Zone()
 		off := float64(offset)
 		r := netstorage.Result{
 			MetricName: metricNameExpected,
@@ -240,10 +242,12 @@ func TestExecSuccess(t *testing.T) {
 	t.Run("timezone_offset(Local)", func(t *testing.T) {
 		t.Parallel()
 		q := `timezone_offset("Local")`
-		offset, err := getTimezoneOffset("Local")
+		loc, err := time.LoadLocation("Local")
 		if err != nil {
 			t.Fatalf("cannot obtain timezone: %s", err)
 		}
+		at := time.Unix(timestampsExpected[0]/1000, 0)
+		_, offset := at.In(loc).Zone()
 		off := float64(offset)
 		r := netstorage.Result{
 			MetricName: metricNameExpected,

app/vmselect/promql/transform.go

@@ -2178,21 +2178,22 @@ func transformTimezoneOffset(tfa *transformFuncArg) ([]*timeseries, error) {
 	if err != nil {
 		return nil, fmt.Errorf("cannot get timezone name: %w", err)
 	}
-	tzOffset, err := getTimezoneOffset(tzString)
-	if err != nil {
-		return nil, fmt.Errorf("cannot get timezone offset for %q: %w", tzString, err)
-	}
-	rv := evalNumber(tfa.ec, float64(tzOffset))
-	return rv, nil
-}
-
-func getTimezoneOffset(tzString string) (int, error) {
 	loc, err := time.LoadLocation(tzString)
 	if err != nil {
-		return 0, fmt.Errorf("cannot load timezone %q: %w", tzString, err)
+		return nil, fmt.Errorf("cannot load timezone %q: %w", tzString, err)
 	}
-	_, tzOffset := time.Now().In(loc).Zone()
+
-	return tzOffset, nil
+	var ts timeseries
+	ts.denyReuse = true
+	timestamps := tfa.ec.getSharedTimestamps()
+	values := make([]float64, len(timestamps))
+	for i, v := range timestamps {
+		_, offset := time.Unix(v/1000, 0).In(loc).Zone()
+		values[i] = float64(offset)
+	}
+	ts.Values = values
+	ts.Timestamps = timestamps
+	return []*timeseries{&ts}, nil
 }
 
 func transformTime(tfa *transformFuncArg) ([]*timeseries, error) {

docs/CHANGELOG.md

@@ -58,6 +58,7 @@ scrape_configs:
 * BUGFIX: limit max memory occupied by the cache, which stores parsed regular expressions. Previously too long regular expressions passed in [MetricsQL queries](https://docs.victoriametrics.com/MetricsQL.html) could result in big amounts of used memory (e.g. multiple of gigabytes). Now the max cache size for parsed regexps is limited to a a few megabytes.
 * BUGFIX: [MetricsQL](https://docs.victoriametrics.com/MetricsQL.html): properly handle partial counter resets when calculating [rate](https://docs.victoriametrics.com/MetricsQL.html#rate), [irate](https://docs.victoriametrics.com/MetricsQL.html#irate) and [increase](https://docs.victoriametrics.com/MetricsQL.html#increase) functions. Previously these functions could return zero values after partial counter resets until the counter increases to the last value before partial counter reset. See [this issue](https://github.com/VictoriaMetrics/VictoriaMetrics/issues/2787).
 * BUGFIX: [MetricsQL](https://docs.victoriametrics.com/MetricsQL.html): properly calculate [histogram_quantile](https://docs.victoriametrics.com/MetricsQL.html#histogram_quantile) over Prometheus buckets with unexpected values. See [this issue](https://github.com/VictoriaMetrics/VictoriaMetrics/issues/2819).
+* BUGFIX: [MetricsQL](https://docs.victoriametrics.com/MetricsQL.html): properly evaluate [timezone_offset](https://docs.victoriametrics.com/MetricsQL.html#timezone_offset) function over time range covering time zone offset switches. See [this issue](https://github.com/VictoriaMetrics/VictoriaMetrics/issues/2771).
 * BUGFIX: [vmagent](https://docs.victoriametrics.com/vmagent.html): properly add service-level labels (`__meta_kubernetes_service_*`) to discovered targets for `role: endpointslice` in [kubernetes_sd_config](https://prometheus.io/docs/prometheus/latest/configuration/configuration/#kubernetes_sd_config). Previously these labels were missing. See [this issue](https://github.com/VictoriaMetrics/VictoriaMetrics/issues/2823).
 * BUGFIX: [vmagent](https://docs.victoriametrics.com/vmagent.html): make sure that [stale markers](https://docs.victoriametrics.com/vmagent.html#prometheus-staleness-markers) are generated with the actual timestamp when unsuccessful scrape occurs. This should prevent from possible time series overlap on scrape target restart in dynmaic envirnoments such as Kubernetes.
 * BUGFIX: [vmagent](https://docs.victoriametrics.com/vmagent.html): properly reload changed `-promscrape.config` file when `-promscrape.configCheckInterval` option is set. The changed config file wasn't reloaded in this case since [v1.69.0](#v1690). See [this pull request](https://github.com/VictoriaMetrics/VictoriaMetrics/pull/2786). Thanks to @ttyv for the fix.