Add non-collectd load monitor and make default (#1140)

The old collectd/load monitor is still present for backwards compatibility

The metrics are exactly the same except for the lack of a 'plugin' dimension

Author: benkeith-splunk

deployments/ansible/README.md

@@ -38,7 +38,7 @@ This role sources the following variables:
         - type: collectd/df
         - type: disk
         - type: collectd/interface
-        - type: collectd/load
+        - type: load
         - type: collectd/memory
         - type: collectd/vmem
         - type: collectd/signalfx-metadata

deployments/ansible/example-config.yml

@@ -7,7 +7,7 @@ sfx_agent_config:
     - type: collectd/df
     - type: disk
     - type: collectd/interface
-    - type: collectd/load
+    - type: load
     - type: collectd/memory
     - type: collectd/signalfx-metadata
       omitProcessInfo: true

deployments/chef/README.md

@@ -53,7 +53,7 @@ node['signalfx_agent']['conf'] = {
     {type: "collectd/df"},
     {type: "disk"},
     {type: "collectd/interface"},
-    {type: "collectd/load"},
+    {type: "load"},
     {type: "collectd/memory"},
     {"type": "collectd/signalfx-metadata", "omitProcessInfo": true},
     {type: "collectd/vmem"}

deployments/chef/example_attrs.json

@@ -12,7 +12,7 @@
       {"type": "collectd/df"},
       {"type": "disk"},
       {"type": "collectd/interface"},
-      {"type": "collectd/load"},
+      {"type": "load"},
       {"type": "collectd/memory"},
       {"type": "collectd/signalfx-metadata", "omitProcessInfo": true},
       {"type": "collectd/vmem"},

deployments/docker/agent.yaml

@@ -27,7 +27,7 @@ monitors:
     hostFSPath: /hostfs
   - type: disk
   - type: collectd/interface
-  - type: collectd/load
+  - type: load
   - type: collectd/memory
   - type: collectd/vmem
   - type: collectd/signalfx-metadata

deployments/ecs/agent.yaml

@@ -41,7 +41,7 @@ monitors:
     hostFSPath: /hostfs
   - type: disk
   - type: collectd/interface
-  - type: collectd/load
+  - type: load
   - type: collectd/memory
   - type: collectd/signalfx-metadata
     omitProcessInfo: true

deployments/fargate/agent.yaml

@@ -21,7 +21,7 @@ monitors:
   - type: collectd/cpu
   - type: disk
   - type: collectd/interface
-  - type: collectd/load
+  - type: load
   - type: collectd/memory
   - type: collectd/protocols
   - type: collectd/signalfx-metadata

deployments/k8s/configmap.yaml

@@ -37,7 +37,7 @@ data:
       hostFSPath: /hostfs
     - type: disk
     - type: collectd/interface
-    - type: collectd/load
+    - type: load
     - type: collectd/memory
     - type: collectd/protocols
     - type: collectd/signalfx-metadata

deployments/k8s/helm/signalfx-agent/templates/configmap.yaml

@@ -61,7 +61,7 @@ data:
       hostFSPath: {{ .Values.hostFSPath }}
     - type: disk
     - type: collectd/interface
-    - type: collectd/load
+    - type: load
     - type: collectd/memory
     - type: collectd/protocols
     - type: collectd/signalfx-metadata

deployments/puppet/README.md

@@ -22,7 +22,7 @@ class accepts the following parameters:
         {type: "collectd/df"},
         {type: "disk"},
         {type: "collectd/interface"},
-        {type: "collectd/load"},
+        {type: "load"},
         {type: "collectd/memory"},
         {type: "collectd/protocols"},
         {type: "collectd/signalfx-metadata", "omitProcessInfo": true},

deployments/salt/README.md

@@ -49,7 +49,7 @@ signalfx-agent:
       - type: collectd/df
       - type: disk
       - type: collectd/interface
-      - type: collectd/load
+      - type: load
       - type: collectd/memory
       - type: collectd/vmem
       - type: collectd/signalfx-metadata

deployments/salt/pillar.example

@@ -11,7 +11,7 @@ signalfx-agent:
       - type: collectd/df
       - type: disk
       - type: collectd/interface
-      - type: collectd/load
+      - type: load
       - type: collectd/memory
       - type: collectd/signalfx-metadata
         omitProcessInfo: true

docs/monitor-config.md

@@ -126,6 +126,7 @@ These are all of the monitors included in the agent, along with their possible c
 - [kubernetes-proxy](./monitors/kubernetes-proxy.md)
 - [kubernetes-scheduler](./monitors/kubernetes-scheduler.md)
 - [kubernetes-volumes](./monitors/kubernetes-volumes.md)
+- [load](./monitors/load.md)
 - [logstash](./monitors/logstash.md)
 - [logstash-tcp](./monitors/logstash-tcp.md)
 - [memory](./monitors/memory.md)

docs/monitors/collectd-load.md

@@ -13,6 +13,10 @@ Monitor Type: `collectd/load` ([Source](https://github.com/signalfx/signalfx-age
 Monitors process load on the host using the collectd
 [Load plugin](https://collectd.org/wiki/index.php/Plugin:Load).
 
+**This monitor has been deprecated in favor of the `load` monitor. That
+monitor emits the same metrics and is fully compatible.  This
+`collectd/load` monitor will be removed in a future release of the agent."
+
 
 ## Configuration
 

docs/monitors/load.md

@@ -0,0 +1,76 @@
+<!--- GENERATED BY gomplate from scripts/docs/templates/monitor-page.md.tmpl --->
+
+# load
+
+Monitor Type: `load` ([Source](https://github.com/signalfx/signalfx-agent/tree/master/pkg/monitors/load))
+
+**Accepts Endpoints**: No
+
+**Multiple Instances Allowed**: **No**
+
+## Overview
+
+Monitors process load on the host. Process load is the average number of
+running or waiting processes over a certain time period (1, 5, and 15
+minutes).
+
+This monitor is only available on Linux.
+
+
+## Configuration
+
+To activate this monitor in the Smart Agent, add the following to your
+agent config:
+
+```
+monitors:  # All monitor config goes under this key
+ - type: load
+   ...  # Additional config
+```
+
+**For a list of monitor options that are common to all monitors, see [Common
+Configuration](../monitor-config.md#common-configuration).**
+
+
+This monitor has no configuration options.
+## Metrics
+
+These are the metrics available for this monitor.
+Metrics that are categorized as
+[container/host](https://docs.signalfx.com/en/latest/admin-guide/usage.html#about-custom-bundled-and-high-resolution-metrics)
+(*default*) are ***in bold and italics*** in the list below.
+
+
+ - ***`load.longterm`*** (*gauge*)<br>    Average CPU load per core over the last 15 minutes
+ - ***`load.midterm`*** (*gauge*)<br>    Average CPU load per core over the last five minutes
+ - ***`load.shortterm`*** (*gauge*)<br>    Average CPU load per core over the last one minute
+
+### Non-default metrics (version 4.7.0+)
+
+**The following information applies to the agent version 4.7.0+ that has
+`enableBuiltInFiltering: true` set on the top level of the agent config.**
+
+To emit metrics that are not _default_, you can add those metrics in the
+generic monitor-level `extraMetrics` config option.  Metrics that are derived
+from specific configuration options that do not appear in the above list of
+metrics do not need to be added to `extraMetrics`.
+
+To see a list of metrics that will be emitted you can run `agent-status
+monitors` after configuring this monitor in a running agent instance.
+
+### Legacy non-default metrics (version < 4.7.0)
+
+**The following information only applies to agent version older than 4.7.0. If
+you have a newer agent and have set `enableBuiltInFiltering: true` at the top
+level of your agent config, see the section above. See upgrade instructions in
+[Old-style whitelist filtering](../legacy-filtering.md#old-style-whitelist-filtering).**
+
+If you have a reference to the `whitelist.json` in your agent's top-level
+`metricsToExclude` config option, and you want to emit metrics that are not in
+that whitelist, then you need to add an item to the top-level
+`metricsToInclude` config option to override that whitelist (see [Inclusion
+filtering](../legacy-filtering.md#inclusion-filtering).  Or you can just
+copy the whitelist.json, modify it, and reference that in `metricsToExclude`.
+
+
+

packaging/etc/agent.yaml

@@ -24,7 +24,7 @@ monitors:
   - type: collectd/df
   - type: disk
   - type: collectd/interface
-  - type: collectd/load
+  - type: load
   - type: collectd/memory
   - type: collectd/signalfx-metadata
     omitProcessInfo: true

pkg/core/modules.go

@@ -45,6 +45,7 @@ import (
 	_ "github.com/signalfx/signalfx-agent/pkg/monitors/jaegergrpc"
 	_ "github.com/signalfx/signalfx-agent/pkg/monitors/jmx"
 	_ "github.com/signalfx/signalfx-agent/pkg/monitors/kubernetes"
+	_ "github.com/signalfx/signalfx-agent/pkg/monitors/load"
 	_ "github.com/signalfx/signalfx-agent/pkg/monitors/logstash/logstash"
 	_ "github.com/signalfx/signalfx-agent/pkg/monitors/logstash/tcp"
 	_ "github.com/signalfx/signalfx-agent/pkg/monitors/memory"

pkg/monitors/collectd/load/metadata.yaml

@@ -3,6 +3,10 @@ monitors:
   doc: |
     Monitors process load on the host using the collectd
     [Load plugin](https://collectd.org/wiki/index.php/Plugin:Load).
+
+    **This monitor has been deprecated in favor of the `load` monitor. That
+    monitor emits the same metrics and is fully compatible.  This
+    `collectd/load` monitor will be removed in a future release of the agent."
   metrics:
     load.longterm:
       description: Average CPU load per core over the last 15 minutes

pkg/monitors/load/genmetadata.go

@@ -0,0 +1,70 @@
+package load
+
+import (
+	"context"
+	"runtime"
+	"time"
+
+	"github.com/shirou/gopsutil/load"
+	"github.com/signalfx/golib/v3/datapoint"
+	"github.com/signalfx/signalfx-agent/pkg/core/config"
+	"github.com/signalfx/signalfx-agent/pkg/monitors"
+	"github.com/signalfx/signalfx-agent/pkg/monitors/types"
+	"github.com/signalfx/signalfx-agent/pkg/utils"
+	"github.com/sirupsen/logrus"
+	log "github.com/sirupsen/logrus"
+)
+
+func init() {
+	if runtime.GOOS != "windows" {
+		monitors.Register(&monitorMetadata, func() interface{} { return &Monitor{} }, &Config{})
+	}
+}
+
+// Config for this monitor
+type Config struct {
+	config.MonitorConfig `singleInstance:"true" acceptsEndpoints:"false"`
+}
+
+// Monitor for load
+type Monitor struct {
+	Output types.Output
+	cancel func()
+	logger logrus.FieldLogger
+}
+
+// Configure is the main function of the monitor, it will report host metadata
+// on a varied interval
+func (m *Monitor) Configure(conf *Config) error {
+	m.logger = logrus.WithFields(log.Fields{"monitorType": monitorType})
+
+	// create contexts for managing the the plugin loop
+	var ctx context.Context
+	ctx, m.cancel = context.WithCancel(context.Background())
+
+	// gather metrics on the specified interval
+	utils.RunOnInterval(ctx, func() {
+		avgLoad, err := load.Avg()
+		if err != nil {
+			m.logger.WithError(err).Error("Failed to get load statistics")
+			return
+		}
+
+		numCPUs := float64(runtime.NumCPU())
+
+		m.Output.SendDatapoints([]*datapoint.Datapoint{
+			datapoint.New(loadLongterm, nil, datapoint.NewFloatValue(avgLoad.Load15/numCPUs), datapoint.Gauge, time.Time{}),
+			datapoint.New(loadMidterm, nil, datapoint.NewFloatValue(avgLoad.Load5/numCPUs), datapoint.Gauge, time.Time{}),
+			datapoint.New(loadShortterm, nil, datapoint.NewFloatValue(avgLoad.Load1/numCPUs), datapoint.Gauge, time.Time{}),
+		}...)
+	}, time.Duration(conf.IntervalSeconds)*time.Second)
+
+	return nil
+}
+
+// Shutdown stops the metric sync
+func (m *Monitor) Shutdown() {
+	if m.cancel != nil {
+		m.cancel()
+	}
+}

pkg/monitors/load/load.go

@@ -0,0 +1,25 @@
+monitors:
+- dimensions:
+  doc: |
+    Monitors process load on the host. Process load is the average number of
+    running or waiting processes over a certain time period (1, 5, and 15
+    minutes).
+
+    This monitor is only available on Linux.
+
+  metrics:
+    load.longterm:
+      description: Average CPU load per core over the last 15 minutes
+      default: true
+      type: gauge
+    load.midterm:
+      description: Average CPU load per core over the last five minutes
+      default: true
+      type: gauge
+    load.shortterm:
+      description: Average CPU load per core over the last one minute
+      default: true
+      type: gauge
+  monitorType: load
+  properties:
+