Network Monitoring Extension

Use Case

Use for monitoring network related metrics.

It uses the machine agent's Sigar library to retrieve the metrics, however these metrics (all or partial) can be overriden through scripting if required.

This extension works only with standalone machine agent.

Note : By default, the Machine agent and AppServer agent can only send a fixed number of metrics to the controller. To change this limit, please follow the instructions mentioned here.

Installation

1. Download NetworkMonitor.zip and unzip it to <machine_agent_dir>/monitors/
2. Edit config.yaml file and provide the required configuration (see Configuration section)
3. Restart the Machine Agent.

Configuration

config.yaml

Note: Please avoid using tab (\t) when editing yaml files. You may want to validate the yaml file using a yaml validator.

Below is an example config for monitoring multiple interfaces with enabled metrics scripting override:

networkInterfaces: [ eth0, eth1
]

overrideMetricsUsingScriptFile: true

scriptTimeoutInSec: 60

scriptFiles:
  - osType: windows
    filePath: monitors/NetworkMonitor/scripts/windows-metrics.bat
  - osType: unixBase
    filePath: monitors/NetworkMonitor/scripts/unix-base-metrics.sh

# Comma separated full path of the metrics for which delta (present - previous) needs to be computed.
# For eg. deltaMetrics: ["TCP|State|Bound", "TCP|Segments Retransmitted"]
deltaMetrics: []

metricPrefix:  "Custom Metrics|Network|"

Script File

There maybe cases where Sigar doesn't work on your environment or you've a preferred way of retrieving the metrics. In any case, you have the flexibility to override some or all metrics using scripting.

There are two script templates provided:

1. unix-base-metrics.sh for non-windows environment
2. windows-metrics.bat for windows environment

Update the relevant script template by uncommenting out the metrics you wish to override and provide the value.

Below is an example override on some TCP State metrics in unix-base-metrics.sh:

...
echo "name=TCP|State|Close Wait,value=" `netstat -an | grep -i CLOSE_WAIT | wc -l`
echo "name=TCP|State|Establised,value=" `netstat -an | grep -i ESTABLISHED | wc -l`
echo "name=TCP|State|Fin Wait1,value=" `netstat -an | grep -i FIN_WAIT_1 | wc -l`
echo "name=TCP|State|Fin Wait2,value=" `netstat -an | grep -i FIN_WAIT_2 | wc -l`
echo "name=TCP|State|Listen,value=" `netstat -an | grep -i LISTEN | wc -l`
...

Below is an example override on custom metrics in windows-metrics.bat:

...
set cmd="netstat -an | find "192.168.31" /c"
FOR /F %%i IN (' %cmd% ') DO SET val=%%i
echo name=Autotrader^|Listen,value=%val%
...

Metrics

Metric path is typically: Application Infrastructure Performance|<Tier>|Custom Metrics|Network| followed by the individual metrics below:

Network Interface

Note: <network_interface> is replaced with the actual network interface name, e.g eth0

TCP

TCP State

Other

Troubleshooting

1. Verify Machine Agent Data:Please start the Machine Agent without the extension and make sure that it reports data. Verify that the machine agent status is UP and it is reporting Hardware Metrics
2. If there is a java.lang.NoClassDefFoundError: org/hyperic/sigar/SigarException in machine-agent.log, please copy all the sigar libraries/dependencies from MachineAgent_Home/lib to MachineAgent_Home/monitorsLibs including sigar.jar and .so files to avoid this error
3. Collect Debug Logs: Edit the file, <MachineAgent>/conf/logging/log4j.xml and update the level of the appender "com.appdynamics"  to debug

Custom Dashboard

Platform Tested

Version

1.0.4 4th Aug, 2016 - Support for configurable delta metrics

Contributing

Always feel free to fork and contribute any changes directly here on GitHub

Support

For any questions or feature request, please contact AppDynamics Center of Excellence.