Merge pull request #1 from ncabatoff/feature-queries_from_file

fix bug with multiple columns in same metric

Author: noushi

.travis.yml

@@ -7,6 +7,7 @@ go:
 script:
 - make all
 - make docker
+- make test-integration
 after_success:
 - if [ "$TRAVIS_BRANCH" == "master" ]; then docker login -e $DOCKER_EMAIL -u $DOCKER_USER
   -p $DOCKER_PASS ; docker push wrouesnel/postgres_exporter ; fi

Makefile

@@ -21,6 +21,9 @@ vet:
 test:
 	go test -v .
 
+test-integration:
+	tests/test-smoke
+
 # Do a self-contained docker build - we pull the official upstream container,
 # then template out a dockerfile which builds the real image.
 docker-build: postgres_exporter

README.md

@@ -65,6 +65,10 @@ for l in StringIO(x):
 Adjust the value of the resultant prometheus value type appropriately. This helps build
 rich self-documenting metrics for the exporter.
 
+### Adding new metrics via a config file
+
+The -extend.query-path command-line argument specifies a YAML file containing additional queries to run.
+Some examples are provided in [queries.yaml](queries.yaml).
 
 ### Running as non-superuser
 

postgres_exporter.go

@@ -1,27 +1,24 @@
 package main
 
 import (
-	//"bytes"
 	"database/sql"
 	"flag"
 	"fmt"
+	"io/ioutil"
+	"math"
 	"net/http"
 	"os"
-	//"regexp"
-	//"strconv"
-	//"strings"
-	"math"
+	"strconv"
 	"time"
-	"io/ioutil"
+
 	"gopkg.in/yaml.v2"
 
 	_ "github.com/lib/pq"
 	"github.com/prometheus/client_golang/prometheus"
-	"github.com/prometheus/log"
-	"strconv"
+	"github.com/prometheus/common/log"
 )
 
-var Version string = "0.0.0-dev"
+var Version string = "0.0.1"
 
 var (
 	listenAddress = flag.String(
@@ -36,6 +33,10 @@ var (
 		"extend.query-path", "",
 		"Path to custom queries to run.",
 	)
+	onlyDumpMaps = flag.Bool(
+		"dumpmaps", false,
+		"Do not run, simply dump the maps.",
+	)
 )
 
 // Metric name parts.
@@ -111,6 +112,21 @@ var variableMaps = map[string]map[string]ColumnMapping{
 	},
 }
 
+func dumpMaps() {
+	for name, cmap := range metricMaps {
+		query, ok := queryOverrides[name]
+		if ok {
+			fmt.Printf("%s: %s\n", name, query)
+		} else {
+			fmt.Println(name)
+		}
+		for column, details := range cmap {
+			fmt.Printf("  %-40s %v\n", column, details)
+		}
+		fmt.Println()
+	}
+}
+
 var metricMaps = map[string]map[string]ColumnMapping{
 	"pg_stat_bgwriter": map[string]ColumnMapping{
 		"checkpoints_timed":     {COUNTER, "Number of scheduled checkpoints that have been performed", nil},
@@ -235,7 +251,6 @@ func addQueries(queriesPath string) (err error) {
 		return err
 	}
 
-
 	for metric, specs := range extra {
 		for key, value := range specs.(map[interface{}]interface{}) {
 			switch key.(string) {
@@ -249,14 +264,16 @@ func addQueries(queriesPath string) (err error) {
 
 					for n, a := range column {
 						var cmap ColumnMapping
-						var metric_map map[string]ColumnMapping
 
-						metric_map = make(map[string]ColumnMapping)
+						metric_map, ok := metricMaps[metric]
+						if !ok {
+							metric_map = make(map[string]ColumnMapping)
+						}
 
 						name := n.(string)
 
 						for attr_key, attr_val := range a.(map[interface{}]interface{}) {
-							switch(attr_key.(string)) {
+							switch attr_key.(string) {
 							case "usage":
 								usage, err := stringToColumnUsage(attr_val.(string))
 								if err != nil {
@@ -355,6 +372,10 @@ func makeDescMap(metricMaps map[string]map[string]ColumnMapping) map[string]Metr
 							return math.NaN(), false
 						}
 
+						if durationString == "-1" {
+							return math.NaN(), false
+						}
+
 						d, err := time.ParseDuration(durationString)
 						if err != nil {
 							log.Errorln("Failed converting result to metric:", columnName, in, err)
@@ -374,7 +395,7 @@ func makeDescMap(metricMaps map[string]map[string]ColumnMapping) map[string]Metr
 
 // convert a string to the corresponding ColumnUsage
 func stringToColumnUsage(s string) (u ColumnUsage, err error) {
-	switch(s) {
+	switch s {
 	case "DISCARD":
 		u = DISCARD
 
@@ -418,6 +439,13 @@ func dbToFloat64(t interface{}) (float64, bool) {
 			return math.NaN(), false
 		}
 		return result, true
+	case string:
+		result, err := strconv.ParseFloat(v, 64)
+		if err != nil {
+			log.Infoln("Could not parse string:", err)
+			return math.NaN(), false
+		}
+		return result, true
 	case nil:
 		return math.NaN(), true
 	default:
@@ -536,7 +564,7 @@ func (e *Exporter) scrape(ch chan<- prometheus.Metric) {
 
 	db, err := sql.Open("postgres", e.dsn)
 	if err != nil {
-		log.Println("Error opening connection to database:", err)
+		log.Infoln("Error opening connection to database:", err)
 		e.error.Set(1)
 		return
 	}
@@ -582,7 +610,7 @@ func (e *Exporter) scrape(ch chan<- prometheus.Metric) {
 			// Don't fail on a bad scrape of one metric
 			rows, err := db.Query(query)
 			if err != nil {
-				log.Println("Error running query on database: ", namespace, err)
+				log.Infoln("Error running query on database: ", namespace, err)
 				e.error.Set(1)
 				return
 			}
@@ -591,7 +619,7 @@ func (e *Exporter) scrape(ch chan<- prometheus.Metric) {
 			var columnNames []string
 			columnNames, err = rows.Columns()
 			if err != nil {
-				log.Println("Error retrieving column list for: ", namespace, err)
+				log.Infoln("Error retrieving column list for: ", namespace, err)
 				e.error.Set(1)
 				return
 			}
@@ -611,7 +639,7 @@ func (e *Exporter) scrape(ch chan<- prometheus.Metric) {
 			for rows.Next() {
 				err = rows.Scan(scanArgs...)
 				if err != nil {
-					log.Println("Error retrieving rows:", namespace, err)
+					log.Infoln("Error retrieving rows:", namespace, err)
 					e.error.Set(1)
 					return
 				}
@@ -666,18 +694,23 @@ func (e *Exporter) scrape(ch chan<- prometheus.Metric) {
 func main() {
 	flag.Parse()
 
-	dsn := os.Getenv("DATA_SOURCE_NAME")
-	if len(dsn) == 0 {
-		log.Fatal("couldn't find environment variable DATA_SOURCE_NAME")
-	}
-
 	if *queriesPath != "" {
 		err := addQueries(*queriesPath)
 		if err != nil {
 			log.Warnln("Unparseable queries file - discarding merge: ", *queriesPath, err)
 		}
 	}
 
+	if *onlyDumpMaps {
+		dumpMaps()
+		return
+	}
+
+	dsn := os.Getenv("DATA_SOURCE_NAME")
+	if len(dsn) == 0 {
+		log.Fatal("couldn't find environment variable DATA_SOURCE_NAME")
+	}
+
 	exporter := NewExporter(dsn)
 	prometheus.MustRegister(exporter)
 

queries.yaml

@@ -1,6 +1,100 @@
 pg_replication:
-  query: 'SELECT EXTRACT(EPOCH FROM (now() - pg_last_xact_replay_timestamp()))::INT as lag'
+  query: "SELECT EXTRACT(EPOCH FROM (now() - pg_last_xact_replay_timestamp()))::INT as lag"
   metrics:
     - lag:
-        usage: 'GAUGE'
-        description: 'Replication lag behind master in seconds'
+        usage: "GAUGE"
+        description: "Replication lag behind master in seconds"
+
+pg_postmaster:
+  query: "SELECT pg_postmaster_start_time as start_time_seconds from pg_postmaster_start_time()"
+  metrics:
+    - start_time_seconds:
+        usage: "GAUGE"
+        description: "Time at which postmaster started"
+
+pg_settings_shared_buffers:
+  query: "SELECT 8192*setting::int as bytes from pg_settings where name = 'shared_buffers'"
+  metrics:
+    - bytes:
+        usage: "GAUGE"
+        description: "Size of shared_buffers"
+
+pg_settings_checkpoint:
+  query: "select (select setting::int from pg_settings where name = 'checkpoint_segments') as segments, (select setting::int from pg_settings where name = 'checkpoint_timeout') as timeout_seconds, (select setting::float from pg_settings where name = 'checkpoint_completion_target') as completion_target"
+  metrics:
+    - segments:
+        usage: "GAUGE"
+        description: "Number of checkpoint segments"
+    - timeout_seconds:
+        usage: "GAUGE"
+        description: "Checkpoint timeout in seconds"
+    - completion_target:
+        usage: "GAUGE"
+        description: "Checkpoint completion target, ranging from 0 to 1"
+
+pg_stat_user_tables:
+  query: "SELECT schemaname, relname, seq_scan, seq_tup_read, idx_scan, idx_tup_fetch, n_tup_ins, n_tup_upd, n_tup_del, n_tup_hot_upd, n_live_tup, n_dead_tup, n_mod_since_analyze, last_vacuum, last_autovacuum, last_analyze, last_autoanalyze, vacuum_count, autovacuum_count, analyze_count, autoanalyze_count FROM pg_stat_user_tables"
+  metrics:
+    - schemaname:
+        usage: "LABEL"
+        description: "Name of the schema that this table is in"
+    - relname:
+        usage: "LABEL"
+        description: "Name of this table"
+    - seq_scan:
+        usage: "COUNTER"
+        description: "Number of sequential scans initiated on this table"
+    - seq_tup_read:
+        usage: "COUNTER"
+        description: "Number of live rows fetched by sequential scans"
+    - idx_scan:
+        usage: "COUNTER"
+        description: "Number of index scans initiated on this table"
+    - idx_tup_fetch:
+        usage: "COUNTER"
+        description: "Number of live rows fetched by index scans"
+    - n_tup_ins:
+        usage: "COUNTER"
+        description: "Number of rows inserted"
+    - n_tup_upd:
+        usage: "COUNTER"
+        description: "Number of rows updated"
+    - n_tup_del:
+        usage: "COUNTER"
+        description: "Number of rows deleted"
+    - n_tup_hot_upd:
+        usage: "COUNTER"
+        description: "Number of rows HOT updated (i.e., with no separate index update required)"
+    - n_live_tup:
+        usage: "GAUGE"
+        description: "Estimated number of live rows"
+    - n_dead_tup:
+        usage: "GAUGE"
+        description: "Estimated number of dead rows"
+    - n_mod_since_analyze:
+        usage: "GAUGE"
+        description: "Estimated number of rows changed since last analyze"
+    - last_vacuum:
+        usage: "GAUGE"
+        description: "Last time at which this table was manually vacuumed (not counting VACUUM FULL)"
+    - last_autovacuum:
+        usage: "GAUGE"
+        description: "Last time at which this table was vacuumed by the autovacuum daemon"
+    - last_analyze:
+        usage: "GAUGE"
+        description: "Last time at which this table was manually analyzed"
+    - last_autoanalyze:
+        usage: "GAUGE"
+        description: "Last time at which this table was analyzed by the autovacuum daemon"
+    - vacuum_count:
+        usage: "COUNTER"
+        description: "Number of times this table has been manually vacuumed (not counting VACUUM FULL)"
+    - autovacuum_count:
+        usage: "COUNTER"
+        description: "Number of times this table has been vacuumed by the autovacuum daemon"
+    - analyze_count:
+        usage: "COUNTER"
+        description: "Number of times this table has been manually analyzed"
+    - autoanalyze_count:
+        usage: "COUNTER"
+        description: "Number of times this table has been analyzed by the autovacuum daemon"