vmsingle and vmagent: keep_input for streaming aggregation

Author: AndrewChubatiuk

api/operator/v1beta1/vmextra_types.go

@@ -593,6 +593,9 @@ type StreamAggrRule struct {
 	// +optional
 	IgnoreOldSamples *bool `json:"ignore_old_samples,omitempty" yaml:"ignore_old_samples,omitempty"`
 
+	// KeepInput instructs whether to keep all the matched input samples after the aggregation
+        KeepInput *bool `json:"keep_input,omitempty" yaml:"keep_input,omitempty"`
+
 	// By is an optional list of labels for grouping input series.
 	//
 	// See also Without.

api/operator/v1beta1/zz_generated.deepcopy.go

@@ -6932,6 +6932,10 @@ spec:
                               interval:
                                 description: Interval is the interval between aggregations.
                                 type: string
+                              keep_input:
+                                description: KeepInput instructs whether to keep all
+                                  the matched input samples after the aggregation
+                                type: boolean
                               keep_metric_names:
                                 description: KeepMetricNames instructs to leave metric
                                   names as is for the output time series without adding
@@ -8720,6 +8724,10 @@ spec:
                         interval:
                           description: Interval is the interval between aggregations.
                           type: string
+                        keep_input:
+                          description: KeepInput instructs whether to keep all the
+                            matched input samples after the aggregation
+                          type: boolean
                         keep_metric_names:
                           description: KeepMetricNames instructs to leave metric names
                             as is for the output time series without adding any suffix.
@@ -34826,6 +34834,10 @@ spec:
                         interval:
                           description: Interval is the interval between aggregations.
                           type: string
+                        keep_input:
+                          description: KeepInput instructs whether to keep all the
+                            matched input samples after the aggregation
+                          type: boolean
                         keep_metric_names:
                           description: KeepMetricNames instructs to leave metric names
                             as is for the output time series without adding any suffix.

config/crd/overlay/crd.yaml

@@ -35,7 +35,7 @@ To perform migration to the `VLSingle` please follow [this docs](https://docs.vi
 * FEATURE: [vmagent](https://docs.victoriametrics.com/operator/resources/vmagent/): added AWS section to VMAgent remoteWrite spec. See [#928](https://github.com/VictoriaMetrics/operator/issues/928).
 * FEATURE: [vmagent](https://docs.victoriametrics.com/operator/resources/vmagent/): introduce global scrape config for VMAgent. See [#1179](https://github.com/VictoriaMetrics/operator/issues/1179).
 * FEATURE: [vmcluster](https://docs.victoriametrics.com/operator/resources/vmcluster/): added the `maxUnavailable` field to VMStorage and VMSelect specs to allow customization of rolling update behavior. See [#1457](https://github.com/VictoriaMetrics/operator/issues/1457) and minimum downtime update strategy [doc](https://docs.victoriametrics.com/victoriametrics/cluster-victoriametrics/#minimum-downtime-strategy). Thanks to the @vpedosyuk for the [1458](https://github.com/VictoriaMetrics/operator/pull/1458) Pull Request
-
+* FEATURE: [vmagent](https://docs.victoriametrics.com/operator/resources/vmagent/) and [vmsingle](https://docs.victoriametrics.com/operator/resources/vmsingle/): adds `keep_input` into `streamAggrConfig`.
 * FEATURE: [vmagent](https://docs.victoriametrics.com/operator/resources/vmagent/): Modify default `remoteWrite.maxDiskUsagePerURL` when statefulStorage.volumeClaimTemplate is set. See [#1209](https://github.com/VictoriaMetrics/operator/issues/1209)
 
 ## [v0.60.2](https://github.com/VictoriaMetrics/operator/releases/tag/v0.60.2)

docs/CHANGELOG.md

@@ -2436,6 +2436,7 @@ Appears in: [StreamAggrConfig](#streamaggrconfig)
 | ignore_old_samples<a href="#streamaggrrule-ignore_old_samples" id="streamaggrrule-ignore_old_samples">#</a><br/>_boolean_ | _(Optional)_<br/>IgnoreOldSamples instructs to ignore samples with old timestamps outside the current aggregation interval. |
 | input_relabel_configs<a href="#streamaggrrule-input_relabel_configs" id="streamaggrrule-input_relabel_configs">#</a><br/>_[RelabelConfig](#relabelconfig) array_ | _(Optional)_<br/>InputRelabelConfigs is an optional relabeling rules, which are applied on the input<br />before aggregation. |
 | interval<a href="#streamaggrrule-interval" id="streamaggrrule-interval">#</a><br/>_string_ | _(Required)_<br/>Interval is the interval between aggregations. |
+| keep_input<a href="#streamaggrrule-keep_input" id="streamaggrrule-keep_input">#</a><br/>_boolean_ | _(Required)_<br/>KeepInput instructs whether to keep all the matched input samples after the aggregation |
 | keep_metric_names<a href="#streamaggrrule-keep_metric_names" id="streamaggrrule-keep_metric_names">#</a><br/>_boolean_ | _(Optional)_<br/>KeepMetricNames instructs to leave metric names as is for the output time series without adding any suffix. |
 | match<a href="#streamaggrrule-match" id="streamaggrrule-match">#</a><br/>_[StringOrArray](#stringorarray)_ | _(Optional)_<br/>Match is a label selector (or list of label selectors) for filtering time series for the given selector.<br /><br />If the match isn't set, then all the input time series are processed. |
 | no_align_flush_to_interval<a href="#streamaggrrule-no_align_flush_to_interval" id="streamaggrrule-no_align_flush_to_interval">#</a><br/>_boolean_ | _(Optional)_<br/>NoAlignFlushToInterval disables aligning of flushes to multiples of Interval.<br />By default flushes are aligned to Interval. |