How to Use the APM Utility and CLI Commands

BBE Event Collection and Visualization is a cloud-based centralized utility that provides a way to capture APM logs that span the life-cycle of APM micro-services. If you have BBE Event Collection and Visualization setup, during the APM setup, you can point BBE Event Collection and Visualization to perform the logging. BBE Event Collection and Visualization has a web-based interface to OpenSearch’s capabilities for advanced searching, aggregation, viewing, and data analysis of collected syslog events.

Use the apm logs utility script to display the logs of events that occur while APM is running. The event logs include events such as those shown in the following non-exhaustive list:

• Pool-domain registration events

• Address allocation failures

• Network entity connection failures

• Startup messages

• Network entity resynchronization events

• Pool and partition utilization threshold and depletion warnings

By default, APM sends logs to the standard output (stdout) of the service. The output displays the circular buffer of all services or of a specified service. You can also enable logging to follow the log output of the running services. Following the log output creates an open session that continuously streams the logs to stdout.

The APM logging functions mask the underlying complexities of the kubectl log command that is collecting the log information. You can still use the kubectl log command, but that is outside the scope of this documentation.

Note:

You can use third-party applications to capture and redirect the stdout stream for the container. Refer to your third-party documentation for assistance. You can also configure Docker with different logging drivers to redirect stdout. Refer to your Docker documentation for assistance.

To display APM logs for all services:

Best Practice:

Use the apm logs ––services option only when you are troubleshooting under the guidance of a Juniper Networks support representative.

To follow the logs for all services, specify the follow option (-f):

To view logs from previous instances of the containers in a Kubernetes pod, specify the previous container option (-p):

You can use standard Ubuntu conventions to redirect the logs to a file or to the terminal and to a file. Refer to the Ubuntu documentation for more information, but you can use the following examples as a starting point:

• Redirect all logs to only a file.

• Redirect all logs to both the screen and to a file.

By default, the utility command reports logs of all severity levels. You can limit the number of logs that are reported by setting a severity level when you start APM services. Include the --log-level option when you start services to specify the minimum severity level of logs that are reported when you also use the logs command. Logs are reported for the specified level and all levels that are more severe. The default level is debug, which is the lowest severity level. This means that by default logs are reported for all levels.

When you set a severity level, it applies to all services.

• To set the severity level for all services:

Best Practice:

Use the ––services option only for troubleshooting under the guidance of a Juniper Networks support representative.

You can specify any of the following severity levels, in order of increasing severity:

Table 3: Severity Level

Severity Level	Description
debug	Detailed information that is typically of interest only when you are trying to diagnose a problem. These logs are often very frequent.
info	Events or non-error conditions of interest. Logs at this level provide confirmation that everything is working as expected. These logs are generally not very frequent.
warning	Indicates that something unexpected happened or that some problem might occur in the near future. A simple example of the latter is the disk space low warning that indicates that you might run out of disk space soon. In either case, the software is still working as expected, but you might want to monitor it more closely. These logs are generally not very frequent.
error	Indicates that a more serious problem has prevented the software from performing some function, but the software has handled the problem as gracefully as possible to continue functioning.
critical	A serious error that indicates that the program itself might be unable to continue running.

You can use the --logset option to display logs either for only APM services or for only prebuilt services. If you do not use this option, then only the APM services logs are displayed.

To display logs for only APM services:

To display logs for prebuilt services:

Prebuilt services are services borrowed from other sources to provide infrastructure functions for APM. These sources include cMGD, haproxy, redis, and redis-sentinel. For example, redis provides database and messaging services, cMGD provides a configuration/CLI service, and so on.

To access the CLI prompt, enter the following apm utility script command:

Enter a question mark to see the available top-level CLI commands. This list of commands is a subset of the Junos OS top-level commands.

For an overview of Junos OS CLI basics, see Day One: Exploring the Junos CLI. For more detailed information, see the CLI User Guide.

To monitor APM, view APM configuration and statistics, or run certain operations manually: