From dae5d782a63f73f82ed650cc07b403f6ae03b3c5 Mon Sep 17 00:00:00 2001
From: Thomas Gummerer <t.gummerer@gmail.com>
Date: Mon, 27 Jan 2020 14:50:34 +0000
Subject: [PATCH] allow setting granularity for summary metrics

The Go client for prometheus aggregates summary metrics over 10
minutes by default, in 5 buckets.  This is not always the behaviour we
want.

Allow tweaking those settings in `statsd_exporter`, so we can
aggregate summary metrics over more or less time, with more or fewer
buckets, and set the cap for the bucket as well.

Signed-off-by: Thomas Gummerer <t.gummerer@gmail.com>
---
 README.md                 | 12 ++++++++++-
 pkg/mapper/mapper.go      |  9 ++++----
 pkg/mapper/mapper_test.go | 45 +++++++++++++++++++++++++++++++++++++++
 registry.go               |  3 +++
 4 files changed, 63 insertions(+), 6 deletions(-)

diff --git a/README.md b/README.md
index 76aa73e..830b57c 100644
--- a/README.md
+++ b/README.md
@@ -236,7 +236,8 @@ mappings:
 
 By default, statsd timers are represented as a Prometheus summary with
 quantiles. You may optionally configure the [quantiles and acceptable
-error](https://prometheus.io/docs/practices/histograms/#quantiles):
+error](https://prometheus.io/docs/practices/histograms/#quantiles), as
+well as adjusting how the summary metric is aggregated:
 
 ```yaml
 mappings:
@@ -257,10 +258,19 @@ mappings:
         error: 0.05
       - quantile: 0.5
         error: 0.005
+    max_summary_age: 30s
+    summary_age_buckets: 3
+    stream_buffer_size: 1000
 ```
 
 The default quantiles are 0.99, 0.9, and 0.5.
 
+The default summary age is 10 minutes, the default number of buckets
+is 5 and the default buffer size is 500.  See also the
+[`golang_client` docs](https://godoc.org/github.com/prometheus/client_golang/prometheus#SummaryOpts).
+The `max_summary_age` corresponds to `SummaryOptions.MaxAge`, `summary_age_buckets`
+to `SummaryOptions.AgeBuckets` and `stream_buffer_size` to `SummaryOptions.BufCap`.
+
 In the configuration, one may also set the timer type to "histogram". The
 default is "summary" as in the plain text configuration format.  For example,
 to set the timer type for a single metric:
diff --git a/pkg/mapper/mapper.go b/pkg/mapper/mapper.go
index e1ec937..61814ad 100644
--- a/pkg/mapper/mapper.go
+++ b/pkg/mapper/mapper.go
@@ -77,7 +77,10 @@ type MetricMapping struct {
 }
 
 type SummaryOptions struct {
-	Quantiles []metricObjective `yaml:"quantiles"`
+	Quantiles  []metricObjective `yaml:"quantiles"`
+	MaxAge     time.Duration     `yaml:"max_age"`
+	AgeBuckets uint32            `yaml:"age_buckets"`
+	BufCap     uint32            `yaml:"buf_cap"`
 }
 
 type HistogramOptions struct {
@@ -226,10 +229,6 @@ func (m *MetricMapper) InitFromYAMLString(fileContents string, cacheSize int) er
 			if currentMapping.SummaryOptions == nil {
 				currentMapping.SummaryOptions = &SummaryOptions{}
 			}
-		}
-
-		if currentMapping.LegacyQuantiles == nil || len(currentMapping.LegacyQuantiles) == 0 {
-			currentMapping.LegacyQuantiles = n.Defaults.Quantiles
 			if currentMapping.LegacyQuantiles != nil && len(currentMapping.LegacyQuantiles) != 0 {
 				currentMapping.SummaryOptions.Quantiles = currentMapping.LegacyQuantiles
 			}
diff --git a/pkg/mapper/mapper_test.go b/pkg/mapper/mapper_test.go
index fbdeb41..dc8a45d 100644
--- a/pkg/mapper/mapper_test.go
+++ b/pkg/mapper/mapper_test.go
@@ -26,6 +26,9 @@ type mappings []struct {
 	notPresent   bool
 	ttl          time.Duration
 	metricType   MetricType
+	maxAge       time.Duration
+	ageBuckets   uint32
+	bufCap       uint32
 }
 
 func TestMetricMapperYAML(t *testing.T) {
@@ -523,6 +526,39 @@ mappings:
 				},
 			},
 		},
+		// Config with summary configuration.
+		{
+			config: `---
+mappings:
+- match: test.*.*
+  timer_type: summary
+  name: "foo"
+  labels: {}
+  summary_options:
+    quantiles:
+      - quantile: 0.42
+        error: 0.04
+      - quantile: 0.7
+        error: 0.002
+    max_age: 5m
+    age_buckets: 2
+    buf_cap: 1000
+  `,
+			mappings: mappings{
+				{
+					statsdMetric: "test.*.*",
+					name:         "foo",
+					labels:       map[string]string{},
+					quantiles: []metricObjective{
+						{Quantile: 0.42, Error: 0.04},
+						{Quantile: 0.7, Error: 0.002},
+					},
+					maxAge:     5 * time.Minute,
+					ageBuckets: 2,
+					bufCap:     1000,
+				},
+			},
+		},
 		// duplicate quantiles are bad
 		{
 			config: `---
@@ -834,6 +870,15 @@ mappings:
 					}
 				}
 			}
+			if mapping.maxAge != 0 && mapping.maxAge != m.SummaryOptions.MaxAge {
+				t.Fatalf("%d.%q: Expected max age %v, got %v", i, metric, mapping.maxAge, m.SummaryOptions.MaxAge)
+			}
+			if mapping.ageBuckets != 0 && mapping.ageBuckets != m.SummaryOptions.AgeBuckets {
+				t.Fatalf("%d.%q: Expected max age %v, got %v", i, metric, mapping.ageBuckets, m.SummaryOptions.AgeBuckets)
+			}
+			if mapping.bufCap != 0 && mapping.bufCap != m.SummaryOptions.BufCap {
+				t.Fatalf("%d.%q: Expected max age %v, got %v", i, metric, mapping.bufCap, m.SummaryOptions.BufCap)
+			}
 		}
 	}
 }
diff --git a/registry.go b/registry.go
index fb47d25..d29dbce 100644
--- a/registry.go
+++ b/registry.go
@@ -344,6 +344,9 @@ func (r *registry) getSummary(metricName string, labels prometheus.Labels, help
 			Name:       metricName,
 			Help:       help,
 			Objectives: objectives,
+			MaxAge:     summaryOptions.MaxAge,
+			AgeBuckets: summaryOptions.AgeBuckets,
+			BufCap:     summaryOptions.BufCap,
 		}, labelNames)
 
 		if err := prometheus.Register(uncheckedCollector{summaryVec}); err != nil {