PMM-5085 Last tweaks.

README.md

@@ -74,3 +74,10 @@ scrape_configs:
 ```
 
 `honor_labels: true` is important because exporter returns metrics with `instance` label set.
+
+## Metrics
+
+Exporter synthesizes [node_exporter](https://github.com/prometheus/node_exporter)-like metrics where possible.
+
+You can see a list of basic monitoring metrics [there](https://github.com/percona/rds_exporter/blob/master/basic/testdata/all.txt)
+and a list of enhanced monitoring metrics in text files [there](https://github.com/percona/rds_exporter/tree/master/enhanced/testdata).

basic/metrics.go

@@ -1,4 +1,3 @@
-// Code generated by go generate; DO NOT EDIT.
 package basic
 
 var Metrics = []Metric{
@@ -119,7 +118,7 @@ var Metrics = []Metric{
 	},
 	{
 		cwName:         "EngineUptime",
-		prometheusName: "node_boot_time",
+		prometheusName: "node_boot_time_seconds",
 		prometheusHelp: "EngineUptime",
 	},
 	{
@@ -129,12 +128,12 @@ var Metrics = []Metric{
 	},
 	{
 		cwName:         "FreeStorageSpace",
-		prometheusName: "node_filesystem_free",
+		prometheusName: "node_filesystem_free_bytes",
 		prometheusHelp: "The amount of available storage space. Units: Bytes",
 	},
 	{
 		cwName:         "FreeableMemory",
-		prometheusName: "node_memory_Cached",
+		prometheusName: "node_memory_Cached_bytes",
 		prometheusHelp: "The amount of available random access memory. Units: Bytes",
 	},
 	{

basic/testdata/all.txt

@@ -150,26 +150,26 @@ aws_rds_write_latency_average{instance="autotest-psql-10",region="us-east-1"} 0.
 aws_rds_write_throughput_average{instance="autotest-aurora-psql-11",region="us-west-2"} 606.8671102439675
 aws_rds_write_throughput_average{instance="autotest-mysql-57",region="us-west-2"} 1.7282965432181072e+06
 aws_rds_write_throughput_average{instance="autotest-psql-10",region="us-east-1"} 8942.486209022882
-# HELP node_boot_time EngineUptime
-# TYPE node_boot_time gauge
-node_boot_time{instance="autotest-aurora-mysql-56",region="us-east-1"} 1.576149092e+09
-node_boot_time{instance="autotest-aurora-psql-11",region="us-west-2"} 1.576147548e+09
+# HELP node_boot_time_seconds EngineUptime
+# TYPE node_boot_time_seconds gauge
+node_boot_time_seconds{instance="autotest-aurora-mysql-56",region="us-east-1"} 1.576149092e+09
+node_boot_time_seconds{instance="autotest-aurora-psql-11",region="us-west-2"} 1.576147548e+09
 # HELP node_cpu_average The percentage of CPU utilization. Units: Percent
 # TYPE node_cpu_average gauge
 node_cpu_average{instance="autotest-aurora-mysql-56",region="us-east-1"} 5.67796610169553
 node_cpu_average{instance="autotest-aurora-psql-11",region="us-west-2"} 4
 node_cpu_average{instance="autotest-mysql-57",region="us-west-2"} 2.95081967213145
 node_cpu_average{instance="autotest-psql-10",region="us-east-1"} 2.29508196721326
-# HELP node_filesystem_free The amount of available storage space. Units: Bytes
-# TYPE node_filesystem_free gauge
-node_filesystem_free{instance="autotest-mysql-57",region="us-west-2"} 6.1639360512e+10
-node_filesystem_free{instance="autotest-psql-10",region="us-east-1"} 6.1936873472e+10
-# HELP node_memory_Cached The amount of available random access memory. Units: Bytes
-# TYPE node_memory_Cached gauge
-node_memory_Cached{instance="autotest-aurora-mysql-56",region="us-east-1"} 8.89040896e+08
-node_memory_Cached{instance="autotest-aurora-psql-11",region="us-west-2"} 5.85375744e+08
-node_memory_Cached{instance="autotest-mysql-57",region="us-west-2"} 8.6335488e+07
-node_memory_Cached{instance="autotest-psql-10",region="us-east-1"} 5.24898304e+08
+# HELP node_filesystem_free_bytes The amount of available storage space. Units: Bytes
+# TYPE node_filesystem_free_bytes gauge
+node_filesystem_free_bytes{instance="autotest-mysql-57",region="us-west-2"} 6.1639360512e+10
+node_filesystem_free_bytes{instance="autotest-psql-10",region="us-east-1"} 6.1936873472e+10
+# HELP node_memory_Cached_bytes The amount of available random access memory. Units: Bytes
+# TYPE node_memory_Cached_bytes gauge
+node_memory_Cached_bytes{instance="autotest-aurora-mysql-56",region="us-east-1"} 8.89040896e+08
+node_memory_Cached_bytes{instance="autotest-aurora-psql-11",region="us-west-2"} 5.85375744e+08
+node_memory_Cached_bytes{instance="autotest-mysql-57",region="us-west-2"} 8.6335488e+07
+node_memory_Cached_bytes{instance="autotest-psql-10",region="us-east-1"} 5.24898304e+08
 # HELP rds_exporter_scrape_duration_seconds Time this RDS scrape took, in seconds.
 # TYPE rds_exporter_scrape_duration_seconds gauge
 rds_exporter_scrape_duration_seconds 1.3998536129999999

enhanced/metrics.go

@@ -137,7 +137,7 @@ type processList struct {
 	TGID         int     `json:"tgid"         help:"The thread group identifier, which is a number representing the process ID to which a thread belongs. This identifier is used to group threads from the same process."`
 	VSS          int     `json:"vss"          help:"The amount of virtual memory allocated to the process, in kilobytes."`
 
-	// TODO Handle this.
+	// TODO Handle this: https://jira.percona.com/browse/PMM-5150
 	VMLimit interface{} `json:"vmlimit" help:"-"`
 }
 
@@ -466,6 +466,8 @@ func (m *osMetrics) makePrometheusMetrics(region string, labels map[string]strin
 		float64(m.Timestamp.Unix()),
 	))
 
+	// TODO Parse uptime: https://jira.percona.com/browse/PMM-2131
+
 	res = append(res, prometheus.MustNewConstMetric(
 		prometheus.NewDesc("rdsosmetrics_General_numVCPUs", "The number of virtual CPUs for the DB instance.", nil, constLabels),
 		prometheus.GaugeValue,

enhanced/scraper_test.go

@@ -17,6 +17,26 @@ import (
 	"github.com/percona/rds_exporter/sessions"
 )
 
+func filterMetrics(metrics []*helpers.Metric) []*helpers.Metric {
+	res := make([]*helpers.Metric, 0, len(metrics))
+	processList := make(map[string]struct{})
+
+	for _, m := range metrics {
+		m.Value = 0
+
+		// skip processList metrics that contain process IDs in labels that change too often
+		if strings.Contains(m.Name, "_processList_") {
+			if _, ok := processList[m.Name]; ok {
+				continue
+			}
+			processList[m.Name] = struct{}{}
+		}
+
+		res = append(res, m)
+	}
+	return res
+}
+
 func TestScraper(t *testing.T) {
 	cfg, err := config.Load("../config.tests.yml")
 	require.NoError(t, err)
@@ -36,15 +56,14 @@ func TestScraper(t *testing.T) {
 
 			for _, instance := range instances {
 				// Test that actually received JSON matches expected JSON.
-				// We can't do that directly, so we do it by comparing produced metrics (minus values).
+				// We can't do that directly, so we do it by comparing produced metrics
+				// (minus values and processList metrics).
 
 				instanceName := strings.TrimPrefix(instance.Instance, "autotest-")
 
 				actualMetrics := helpers.ReadMetrics(metrics[instance.ResourceID])
 				sort.Slice(actualMetrics, func(i, j int) bool { return actualMetrics[i].Less(actualMetrics[j]) })
-				for _, m := range actualMetrics {
-					m.Value = 0
-				}
+				actualMetrics = filterMetrics(actualMetrics)
 				actualLines := helpers.Format(helpers.WriteMetrics(actualMetrics))
 
 				if *golden {
@@ -55,9 +74,7 @@ func TestScraper(t *testing.T) {
 				require.NoError(t, err)
 				expectedMetrics := helpers.ReadMetrics(osMetrics.makePrometheusMetrics(instance.Region, nil))
 				sort.Slice(expectedMetrics, func(i, j int) bool { return expectedMetrics[i].Less(expectedMetrics[j]) })
-				for _, m := range expectedMetrics {
-					m.Value = 0
-				}
+				expectedMetrics = filterMetrics(expectedMetrics)
 				expectedLines := helpers.Format(helpers.WriteMetrics(expectedMetrics))
 
 				// compare both to try to avoid go-difflib bug