Set a timeout for calls to input.Gather

Changing the internal behavior around running plugins. Each plugin
will now have it's own goroutine with it's own ticker. This means that a
hung plugin will not block any other plugins. When a plugin is hung, we
will log an error message every interval, letting users know which
plugin is hung.

Currently the input interface does not have any methods for killing a
running Gather call, so there is nothing we can do but log an "ERROR"
and move on. This will give some visibility into the plugin that is
acting up.

closes #1230
fixes #479

agent/accumulator.go

@@ -4,7 +4,6 @@ import (
 	"fmt"
 	"log"
 	"math"
-	"sync"
 	"time"
 
 	"github.com/influxdata/telegraf"
@@ -22,13 +21,13 @@ func NewAccumulator(
 }
 
 type accumulator struct {
-	sync.Mutex
-
 	metrics chan telegraf.Metric
 
 	defaultTags map[string]string
 
 	debug bool
+	// print every point added to the accumulator
+	trace bool
 
 	inputConfig *internal_models.InputConfig
 
@@ -152,7 +151,7 @@ func (ac *accumulator) AddFields(
 		log.Printf("Error adding point [%s]: %s\n", measurement, err.Error())
 		return
 	}
-	if ac.debug {
+	if ac.trace {
 		fmt.Println("> " + m.String())
 	}
 	ac.metrics <- m
@@ -166,6 +165,14 @@ func (ac *accumulator) SetDebug(debug bool) {
 	ac.debug = debug
 }
 
+func (ac *accumulator) Trace() bool {
+	return ac.trace
+}
+
+func (ac *accumulator) SetTrace(trace bool) {
+	ac.trace = trace
+}
+
 func (ac *accumulator) setDefaultTags(tags map[string]string) {
 	ac.defaultTags = tags
 }

agent/agent.go

@@ -102,71 +102,20 @@ func panicRecover(input *internal_models.RunningInput) {
 	}
 }
 
-// gatherParallel runs the inputs that are using the same reporting interval
-// as the telegraf agent.
-func (a *Agent) gatherParallel(metricC chan telegraf.Metric) error {
-	var wg sync.WaitGroup
-
-	start := time.Now()
-	counter := 0
-	jitter := a.Config.Agent.CollectionJitter.Duration.Nanoseconds()
-	for _, input := range a.Config.Inputs {
-		if input.Config.Interval != 0 {
-			continue
-		}
-
-		wg.Add(1)
-		counter++
-		go func(input *internal_models.RunningInput) {
-			defer panicRecover(input)
-			defer wg.Done()
-
-			acc := NewAccumulator(input.Config, metricC)
-			acc.SetDebug(a.Config.Agent.Debug)
-			acc.setDefaultTags(a.Config.Tags)
-
-			if jitter != 0 {
-				nanoSleep := rand.Int63n(jitter)
-				d, err := time.ParseDuration(fmt.Sprintf("%dns", nanoSleep))
-				if err != nil {
-					log.Printf("Jittering collection interval failed for plugin %s",
-						input.Name)
-				} else {
-					time.Sleep(d)
-				}
-			}
-
-			if err := input.Input.Gather(acc); err != nil {
-				log.Printf("Error in input [%s]: %s", input.Name, err)
-			}
-
-		}(input)
-	}
-
-	if counter == 0 {
-		return nil
-	}
-
-	wg.Wait()
-
-	elapsed := time.Since(start)
-	if !a.Config.Agent.Quiet {
-		log.Printf("Gathered metrics, (%s interval), from %d inputs in %s\n",
-			a.Config.Agent.Interval.Duration, counter, elapsed)
-	}
-	return nil
-}
-
-// gatherSeparate runs the inputs that have been configured with their own
+// gatherer runs the inputs that have been configured with their own
 // reporting interval.
-func (a *Agent) gatherSeparate(
+func (a *Agent) gatherer(
 	shutdown chan struct{},
 	input *internal_models.RunningInput,
+	interval time.Duration,
 	metricC chan telegraf.Metric,
 ) error {
 	defer panicRecover(input)
 
-	ticker := time.NewTicker(input.Config.Interval)
+	ticker := time.NewTicker(interval)
+	defer ticker.Stop()
+
+	jitter := a.Config.Agent.CollectionJitter.Duration.Nanoseconds()
 
 	for {
 		var outerr error
@@ -176,14 +125,23 @@ func (a *Agent) gatherSeparate(
 		acc.SetDebug(a.Config.Agent.Debug)
 		acc.setDefaultTags(a.Config.Tags)
 
-		if err := input.Input.Gather(acc); err != nil {
-			log.Printf("Error in input [%s]: %s", input.Name, err)
+		if jitter != 0 {
+			nanoSleep := rand.Int63n(jitter)
+			d, err := time.ParseDuration(fmt.Sprintf("%dns", nanoSleep))
+			if err != nil {
+				log.Printf("Jittering collection interval failed for plugin %s",
+					input.Name)
+			} else {
+				time.Sleep(d)
+			}
 		}
 
+		gatherWithTimeout(shutdown, input, acc, interval)
+
 		elapsed := time.Since(start)
-		if !a.Config.Agent.Quiet {
-			log.Printf("Gathered metrics, (separate %s interval), from %s in %s\n",
-				input.Config.Interval, input.Name, elapsed)
+		if a.Config.Agent.Debug {
+			log.Printf("Input [%s] Gathered metrics, (%s interval) in %s\n",
+				input.Name, interval, elapsed)
 		}
 
 		if outerr != nil {
@@ -199,6 +157,42 @@ func (a *Agent) gatherSeparate(
 	}
 }
 
+// gatherWithTimeout gathers from the given input, with the given timeout.
+//   when the given timeout is reached, gatherWithTimeout logs an error message
+//   but continues waiting for it to return. This is to avoid leaving behind
+//   hung processes, and to prevent re-calling the same hung process over and
+//   over.
+func gatherWithTimeout(
+	shutdown chan struct{},
+	input *internal_models.RunningInput,
+	acc *accumulator,
+	timeout time.Duration,
+) {
+	ticker := time.NewTicker(timeout)
+	defer ticker.Stop()
+	done := make(chan error)
+	go func() {
+		done <- input.Input.Gather(acc)
+	}()
+
+	for {
+		select {
+		case err := <-done:
+			if err != nil {
+				log.Printf("ERROR in input [%s]: %s", input.Name, err)
+			}
+			return
+		case <-ticker.C:
+			log.Printf("ERROR: input [%s] took longer to collect than "+
+				"collection interval (%s)",
+				input.Name, timeout)
+			continue
+		case <-shutdown:
+			return
+		}
+	}
+}
+
 // Test verifies that we can 'Gather' from all inputs with their configured
 // Config struct
 func (a *Agent) Test() error {
@@ -220,7 +214,7 @@ func (a *Agent) Test() error {
 
 	for _, input := range a.Config.Inputs {
 		acc := NewAccumulator(input.Config, metricC)
-		acc.SetDebug(true)
+		acc.SetTrace(true)
 		acc.setDefaultTags(a.Config.Tags)
 
 		fmt.Printf("* Plugin: %s, Collection 1\n", input.Name)
@@ -348,7 +342,6 @@ func (a *Agent) Run(shutdown chan struct{}) error {
 		i := int64(a.Config.Agent.Interval.Duration)
 		time.Sleep(time.Duration(i - (time.Now().UnixNano() % i)))
 	}
-	ticker := time.NewTicker(a.Config.Agent.Interval.Duration)
 
 	wg.Add(1)
 	go func() {
@@ -359,32 +352,21 @@ func (a *Agent) Run(shutdown chan struct{}) error {
 		}
 	}()
 
+	wg.Add(len(a.Config.Inputs))
 	for _, input := range a.Config.Inputs {
-		// Special handling for inputs that have their own collection interval
-		// configured. Default intervals are handled below with gatherParallel
+		interval := a.Config.Agent.Interval.Duration
+		// overwrite global interval if this plugin has it's own.
 		if input.Config.Interval != 0 {
-			wg.Add(1)
-			go func(input *internal_models.RunningInput) {
-				defer wg.Done()
-				if err := a.gatherSeparate(shutdown, input, metricC); err != nil {
-					log.Printf(err.Error())
-				}
-			}(input)
+			interval = input.Config.Interval
 		}
+		go func(in *internal_models.RunningInput, interv time.Duration) {
+			defer wg.Done()
+			if err := a.gatherer(shutdown, in, interv, metricC); err != nil {
+				log.Printf(err.Error())
+			}
+		}(input, interval)
 	}
 
-	defer wg.Wait()
-
-	for {
-		if err := a.gatherParallel(metricC); err != nil {
-			log.Printf(err.Error())
-		}
-
-		select {
-		case <-shutdown:
-			return nil
-		case <-ticker.C:
-			continue
-		}
-	}
+	wg.Wait()
+	return nil
 }