feat(inputs.vsphere): Flag for more lenient behavior when connect fails on startup (#12828)

plugins/inputs/vsphere/README.md

@@ -234,6 +234,12 @@ to use them.
   ## The Historical Interval value must match EXACTLY the interval in the daily
   # "Interval Duration" found on the VCenter server under Configure > General > Statistics > Statistic intervals
   # historical_interval = "5m"
+
+  ## Specifies plugin behavior regarding disconnected servers
+  ## Available choices :
+  ##   - error: telegraf will return an error on startup if one the servers is unreachable
+  ##   - skip: telegraf will skip unreachable servers on both startup and gather
+  # disconnected_servers_behavior = "error"
 ```
 
 NOTE: To disable collection of a specific resource type, simply exclude all

plugins/inputs/vsphere/endpoint.go

@@ -304,7 +304,17 @@ func (e *Endpoint) initalDiscovery(ctx context.Context) {
 func (e *Endpoint) init(ctx context.Context) error {
 	client, err := e.clientFactory.GetClient(ctx)
 	if err != nil {
+		switch e.Parent.DisconnectedServersBehavior {
+		case "error":
 			return err
+		case "ignore":
+			// Ignore the error and postpone the init until next collection cycle
+			e.log.Warnf("Error connecting to vCenter on init: %s", err)
+			return nil
+		default:
+			return fmt.Errorf("%q is not a valid value for disconnected_servers_behavior",
+				e.Parent.DisconnectedServersBehavior)
+		}
 	}
 
 	// Initial load of custom field metadata
@@ -889,6 +899,15 @@ func (e *Endpoint) Close() {
 
 // Collect runs a round of data collections as specified in the configuration.
 func (e *Endpoint) Collect(ctx context.Context, acc telegraf.Accumulator) error {
+	// Connection could have failed on init, so we need to check for a deferred
+	// init request.
+	if !e.initialized {
+		e.log.Debug("Performing deferred init")
+		err := e.init(ctx)
+		if err != nil {
+			return err
+		}
+	}
 	// If we never managed to do a discovery, collection will be a no-op. Therefore,
 	// we need to check that a connection is available, or the collection will
 	// silently fail.

plugins/inputs/vsphere/sample.conf

@@ -193,3 +193,9 @@
   ## The Historical Interval value must match EXACTLY the interval in the daily
   # "Interval Duration" found on the VCenter server under Configure > General > Statistics > Statistic intervals
   # historical_interval = "5m"
+
+  ## Specifies plugin behavior regarding disconnected servers
+  ## Available choices :
+  ##   - error: telegraf will return an error on startup if one the servers is unreachable
+  ##   - skip: telegraf will skip unreachable servers on both startup and gather
+  # disconnected_servers_behavior = "error"

plugins/inputs/vsphere/vsphere.go

@@ -61,6 +61,7 @@ type VSphere struct {
 	UseIntSamples               bool
 	IPAddresses                 []string
 	MetricLookback              int
+	DisconnectedServersBehavior string
 
 	MaxQueryObjects         int
 	MaxQueryMetrics         int
@@ -195,6 +196,7 @@ func init() {
 			ObjectDiscoveryInterval:     config.Duration(time.Second * 300),
 			Timeout:                     config.Duration(time.Second * 60),
 			HistoricalInterval:          config.Duration(time.Second * 300),
+			DisconnectedServersBehavior: "error",
 		}
 	})
 }

plugins/inputs/vsphere/vsphere_test.go

@@ -4,6 +4,7 @@ import (
 	"context"
 	"crypto/tls"
 	"fmt"
+	"net/url"
 	"os"
 	"strings"
 	"testing"
@@ -141,6 +142,7 @@ func defaultVSphere() *VSphere {
 		CollectConcurrency:          1,
 		Separator:                   ".",
 		HistoricalInterval:          config.Duration(time.Second * 300),
+		DisconnectedServersBehavior: "error",
 	}
 }
 
@@ -449,6 +451,22 @@ func TestCollectionNoClusterMetrics(t *testing.T) {
 	testCollection(t, true)
 }
 
+func TestDisconnectedServerBehavior(t *testing.T) {
+	u, err := url.Parse("https://definitely.not.a.valid.host")
+	require.NoError(t, err)
+	v := defaultVSphere()
+	v.DisconnectedServersBehavior = "error"
+	_, err = NewEndpoint(context.Background(), v, u, v.Log)
+	require.Error(t, err)
+	v.DisconnectedServersBehavior = "ignore"
+	_, err = NewEndpoint(context.Background(), v, u, v.Log)
+	require.NoError(t, err)
+	v.DisconnectedServersBehavior = "something else"
+	_, err = NewEndpoint(context.Background(), v, u, v.Log)
+	require.Error(t, err)
+	require.Equal(t, err.Error(), `"something else" is not a valid value for disconnected_servers_behavior`)
+}
+
 func testCollection(t *testing.T, excludeClusters bool) {
 	mustHaveMetrics := map[string]struct{}{
 		"vsphere.vm.cpu":         {},