Merge pull request #512 from karan/dev_net_metrics

Add network interface stats

config/net-cgroup-system-stats-monitor.json

@@ -0,0 +1,55 @@
+{
+  "net": {
+    "metricsConfigs": {
+      "net/rx_bytes": {
+        "displayName": "net/rx_bytes"
+      },
+      "net/rx_packets": {
+        "displayName": "net/rx_packets"
+      },
+      "net/rx_errors": {
+        "displayName": "net/rx_errors"
+      },
+      "net/rx_dropped": {
+        "displayName": "net/rx_dropped"
+      },
+      "net/rx_fifo": {
+        "displayName": "net/rx_fifo"
+      },
+      "net/rx_frame": {
+        "displayName": "net/rx_frame"
+      },
+      "net/rx_compressed": {
+        "displayName": "net/rx_compressed"
+      },
+      "net/rx_multicast": {
+        "displayName": "net/rx_multicast"
+      },
+      "net/tx_bytes": {
+        "displayName": "net/tx_bytes"
+      },
+      "net/tx_packets": {
+        "displayName": "net/tx_packets"
+      },
+      "net/tx_errors": {
+        "displayName": "net/tx_errors"
+      },
+      "net/tx_dropped": {
+        "displayName": "net/tx_dropped"
+      },
+      "net/tx_fifo": {
+        "displayName": "net/tx_fifo"
+      },
+      "net/tx_collisions": {
+        "displayName": "net/tx_collisions"
+      },
+      "net/tx_carrier": {
+        "displayName": "net/tx_carrier"
+      },
+      "net/tx_compressed": {
+        "displayName": "net/tx_compressed"
+      }
+    }
+  },
+  "invokeInterval": "120s"
+}

config/systemd/node-problem-detector-metric-only.service

@@ -10,7 +10,7 @@ ExecStart=/home/kubernetes/bin/node-problem-detector --v=2 --logtostderr --enabl
           --exporter.stackdriver=/home/kubernetes/node-problem-detector/config/exporter/stackdriver-exporter.json \
           --config.system-log-monitor=/home/kubernetes/node-problem-detector/config/kernel-monitor.json,/home/kubernetes/node-problem-detector/config/docker-monitor.json,/home/kubernetes/node-problem-detector/config/systemd-monitor.json \
           --config.custom-plugin-monitor=/home/kubernetes/node-problem-detector/config/kernel-monitor-counter.json,/home/kubernetes/node-problem-detector/config/systemd-monitor-counter.json \
-          --config.system-stats-monitor=/home/kubernetes/node-problem-detector/config/system-stats-monitor.json          
+          --config.system-stats-monitor=/home/kubernetes/node-problem-detector/config/system-stats-monitor.json,/home/kubernetes/node-problem-detector/config/net-cgroup-system-stats-monitor.json
 
 [Install]
 WantedBy=multi-user.target

pkg/exporters/stackdriver/stackdriver_exporter.go

@@ -73,6 +73,22 @@ var NPDMetricToSDMetric = map[metrics.MetricID]string{
 	metrics.SystemProcsRunning:      "kubernetes.io/internal/node/guest/system/procs_running",
 	metrics.SystemProcsBlocked:      "kubernetes.io/internal/node/guest/system/procs_blocked",
 	metrics.SystemInterruptsTotal:   "kubernetes.io/internal/node/guest/system/interrupts_total",
+	metrics.NetDevRxBytes:           "kubernetes.io/internal/node/guest/net/rx_bytes",
+	metrics.NetDevRxPackets:         "kubernetes.io/internal/node/guest/net/rx_packets",
+	metrics.NetDevRxErrors:          "kubernetes.io/internal/node/guest/net/rx_errors",
+	metrics.NetDevRxDropped:         "kubernetes.io/internal/node/guest/net/rx_dropped",
+	metrics.NetDevRxFifo:            "kubernetes.io/internal/node/guest/net/rx_fifo",
+	metrics.NetDevRxFrame:           "kubernetes.io/internal/node/guest/net/rx_frame",
+	metrics.NetDevRxCompressed:      "kubernetes.io/internal/node/guest/net/rx_compressed",
+	metrics.NetDevRxMulticast:       "kubernetes.io/internal/node/guest/net/rx_multicast",
+	metrics.NetDevTxBytes:           "kubernetes.io/internal/node/guest/net/tx_bytes",
+	metrics.NetDevTxPackets:         "kubernetes.io/internal/node/guest/net/tx_packets",
+	metrics.NetDevTxErrors:          "kubernetes.io/internal/node/guest/net/tx_errors",
+	metrics.NetDevTxDropped:         "kubernetes.io/internal/node/guest/net/tx_dropped",
+	metrics.NetDevTxFifo:            "kubernetes.io/internal/node/guest/net/tx_fifo",
+	metrics.NetDevTxCollisions:      "kubernetes.io/internal/node/guest/net/tx_collisions",
+	metrics.NetDevTxCarrier:         "kubernetes.io/internal/node/guest/net/tx_carrier",
+	metrics.NetDevTxCompressed:      "kubernetes.io/internal/node/guest/net/tx_compressed",
 }
 
 func getMetricTypeConversionFunction(customMetricPrefix string) func(*view.View) string {

pkg/systemstatsmonitor/README.md

@@ -77,18 +77,41 @@ Below metrics are collected from `memory` component:
 * `memory_unevictable_used`: [Unevictable memory][/proc doc] usage, in Bytes.
 * `memory_dirty_used`: Dirty pages usage, in Bytes. Memory usage state is reported under the `state` metric label (e.g. `dirty`, `writeback`). `dirty` means the memory is waiting to be written back to disk, and `writeback` means the memory is actively being written back to disk.
 
-### OS features 
+### OS features
 
-The guest OS features such as KTD kernel, GPU support are collected. Below are the OS 
+The guest OS features such as KTD kernel, GPU support are collected. Below are the OS
 features collected:
 
 * `KTD`: Enabled, if KTD feature is enabled on OS
 * `UnifiedCgroupHierarchy`: Enabled, if Unified hierarchy is enabled on OS.
 * `KernelModuleIntegrity`: Enabled, if load pin security is enabled and modules are signed.
 * `GPUSupport`: Enabled, if OS has GPU drivers installed like nvidia.
 * `UnknownModules`: Enabled, if the OS has third party kernel modules installed.
-UnknownModules are derived from the /proc/modules compared with the known-modules.json. 
+UnknownModules are derived from the /proc/modules compared with the known-modules.json.
 
 And an option:
-`knownModulesConfigPath`: The path to the file that contains the known modules(default 
-modules) can be set. By default, the path is set to `known-modules.json`
+`knownModulesConfigPath`: The path to the file that contains the known modules(default
+modules) can be set. By default, the path is set to `known-modules.json`
+
+### IP Stats (Net Dev)
+
+Below metrics are collected from `net` component:
+
+* `net/rx_bytes`: Cumulative count of bytes received.
+* `net/rx_packets`: Cumulative count of packets received.
+* `net/rx_errors`: Cumulative count of receive errors encountered.
+* `net/rx_dropped`: Cumulative count of packets dropped while receiving.
+* `net/rx_fifo`: Cumulative count of FIFO buffer errors.
+* `net/rx_frame`: Cumulative count of packet framing errors.
+* `net/rx_compressed`: Cumulative count of compressed packets received by the device driver.
+* `net/rx_multicast`: Cumulative count of multicast frames received by the device driver.
+* `net/tx_bytes`: Cumulative count of bytes transmitted.
+* `net/tx_packets`: Cumulative count of packets transmitted.
+* `net/tx_errors`: Cumulative count of transmit errors encountered.
+* `net/tx_dropped`: Cumulative count of packets dropped while transmitting.
+* `net/tx_fifo`: Cumulative count of FIFO buffer errors.
+* `net/tx_collisions`: Cumulative count of collisions detected on the interface.
+* `net/tx_carrier`: Cumulative count of carrier losses detected by the device driver.
+* `net/tx_compressed`: Cumulative count of compressed packets transmitted by the device driver.
+
+All of the above have `interface_name` label for the net interface.

pkg/systemstatsmonitor/labels.go

@@ -42,3 +42,6 @@ const osVersionLabel = "os_version"
 
 // osVersionLabel labels the kernel version
 const kernelVersionLabel = "kernel_version"
+
+// interfaceNameLabel labels the network interface name
+const interfaceNameLabel = "interface_name"