Managing host agents
All host agents that report to Instana are displayed on the Agents View. This page gives a quick overview of all agents, their important properties, and general health. Furthermore, each host agent has its own Agent Management View, which displays real-time information about important metrics and allows you to configure the agent in various ways.
Agent management dashboard
To view an Agent Management Dashboard, on any host dashboard click Open Agent Management.
Furthermore, all Agent Management Dashboards can be reached through the Agents View.
The Agent Management Dashboard displays the agent configuration and current runtime metrics. In the management area, configurations can be changed, and operations such as a restart can be run. Further down runtime metrics, and the current agent log are shown.
Changing agent modes
The host agent mode is relevant for host-based licensing, and it is used to decide whether the agent counts as an Infrastructure or an APM agent. The host agent mode can be toggled by using Change Agent Mode
control.
In the Change Agent Mode
dialog, the following options are provided:
APM
, which maps to theAPM
host agent modeInfrastructure
, which maps to theINFRASTRUCTURE
host agent modeDisabled
, which maps to theOFF
host agent mode
An agent that is Disabled
does not report any data and as such is not included in the Managed Virtual Servers (MVS) count.
Changing log level
The log level
defines the level of detail that the agent logs contain.
In the Change Agent Log Level
dialog, the following options are provided:
INFO
, normal level of detailsDEBUG
, extended details for troubleshootingTRACE
, most details for further problem analysis
Updating agents
This button manually triggers the agent auto update as described at Updates of Dynamic Host Agents.
During the update, the host agent cannot provide any data so you might see a short gap.
Resetting agents
Resetting the agent is similar to restarting the agent process, but the current process remains active. It means that OS level watchdogs and service scripts will not see the process identifier (pid
) changing.
During the reset, the host agent cannot provide any data, so you might see a short gap.
Rebooting agents
Rebooting the agent stops the current agent process and create a new one. The process of the used start script remains unchanged, but operating system level watchdogs and service scripts might see a changed process identifier (pid
)
depending on their watch target.
During the restart, the host agent cannot provide any data, which results in a short gap.
Sensor information
To view the list of sensors and host agent components that are used by the agent, click Sensors Info
.
Configuration management
Git-based configuration management can be enabled and configured in the configuration management section.
Details of this feature, and the used nomenclature are described on the Git-based configuration management page.
To support Configuration Management at least agent-bootstrap
version 1.2.11
is required.
To have access to this feature, users need to have a role that with the Configuration of agents
permission. For more information about users, roles, and permissions, see managing users docs.
Initializing configuration management
If no configuration management is active, it can be initialized with the Initialize
button in the Configuration Management
section.
The initialization dialog provides the settings:
Remote Name
, name the Git remote should be set toRemote Branch
, name of the branch to use for updatesRemote URI
, URI of the Git repository
Initialize & Restart
configures the agent according to these settings and restart it to fetch the newest configuration.
Updating configuration management
If configuration management is already active, its details are displayed in the Configuration Management
section. It includes information about the configured remote, like branch name and URI, and the currently used commit,
like hash and commit message.
To update the configuration, the Configuration Management
dialog can be opened by using the Update
button in the section.
Similar to the initialization dialog the following settings can be changed:
Remote Name
, name the Git remote should be set toRemote Branch
, name of the branch to use for updatesRemote URI
, URI of the Git repository
Update & Restart
configures the agent according to these settings and restart it to fetch the newest configuration.
Self monitoring
Not only is the host agent monitoring other software that is running on the host, but it also does constant lightweight self monitoring. Gathered metrics can be used to observe host agent performance and resource consumption:
Agents view
The Agents view, which is located in the main navigation bar under More > Agents, provides an overview of all host agents that report to the Instana tenant unit.
When you use the search bar on the page, the view can be narrowed down to a subset of agents by using Dynamic Focus Queries.
Bulk agent management
To have access to this feature, users need to have a role with the Configuration of agents
permission. For more information about users, roles, and permissions, see managing users docs.
The Agents view allows for quick access to the following features:
- Update All Agents: updates all reporting agents to the latest version.
- Reset All Agents: resets all reporting agents.
- Instana Agent Installation: agent installation instructions for all supported platforms.
Host agent health status
The host agent health monitoring table contains the following information:
- Agent entity link, which links to the agent management dashboard
- Agent version
- Agent boot version
- Origin of installation
- Update Mode
- Mode (
Disabled
,Infrastructure
, orAPM
) - Log level
- JVM name and version
- Reporting status
Reported agents
As with most features in Instana, the host agent health monitoring view strongly depends on the current time configuration.
There are two important things to understand:
- The table presents all agents, which reported within the current time window.
- The status (
reporting
ornot reporting
) is based on the selected time.
The following conceptual image shows how the reporting indicator behaves.
In this example image, the time window was set so that three agents are listed. The end of the selected time window was set somewhere around 6:20am. The agent table would now indicate Agent 1 as not reporting because it stops reporting at 6:00am, so before 6:20am. The other two agents report at that time.
Debugging
Creating a host agent heap dump
To create a heap dump of the host agent, run the following command:
TS=`date +%s` && /opt/instana/agent/jvm/bin/jmap -dump:file=/tmp/agent-dump-$TS.hprof `cat /opt/instana/agent/agent.pid` && gzip /tmp/agent-dump-$TS.hprof
Troubleshooting
There might be cases where your setup does not work at first. If this troubleshooting section does not answer the questions you have, contact the IBM Instana support team with information about your experience, so that we can help you and update our documentation accordingly.
Instana Agent contains vulnerable Log4j library
Monitoring issue type: agent_log4j_vulnerability
Due to a vulnerability discovered in Apache Log4j (CVE-2021-44228 and CVE-2021-45046) we have provided an update to our agent component. It is based on the Pax Logging library as used by our agent packages.
We recommend to update your Instana Agent installation packages to (re-)install the latest Agent version as described here: Installing host agents. Latest packages contain upgraded Pax Logging libraries that mitigate these CVE's. For more information about how Instana mitigates the risk for the Apache Log4j CVE-2021-44228 and CVE-2021-45046, see incident page.
Instana Agent is missing the /tmp
directory
Monitoring issue type: agent_tmp_directory_missing
The Instana host agent creates temporary JAR files in the $TMP/.instana
directory. These JAR files are required for the JVM monitoring. This warning indicates that either the necessary JAR files are missing or the Instana tmp
directory is unavailable, which can lead to Java attachment issues. The Instana Agent periodically monitors the state of the Instana tmp directory and its relevant JAR files to prevent Java attachment failure. To avoid the issue , make sure
that the tmp
location has read and write permissions and avoid unnecessary cleanup of the /tmp
location.
Insufficient disk space for the /tmp
directory
Monitoring issue type: insufficient-disk-space-for-tmp-directory
At run time, the Instana agent container creates temporary JAR files in the $TMP/.instana
directory. These JAR files are required for the JVM monitoring. This warning indicates that insufficient space in the tmp
location
to accommodate temporary files, which can lead to a Java attachment failure. To address this issue, increase the allocated space for tmp
in the container deployment.
Missing TLSv1.3 support for Instana agent
Monitoring issue type: agent_jvm_tls_1_3_missing
The Instana agent requires TLSv1.3
support on the installed host. If TLSv1.3
is not supported, monitoring capabilities will be degraded.
You are recommended to update your Instana agent TLS version to TLSv1.3
. The following TLS versions are enabled: TLSv1
, TLSv1.1
, TLSv1.2
, TLSv1.3
. For more information, see
Setting up TLS encryption for agent endpoint.
Solaris JVM file permission issue
Monitoring issue type: solaris_jspawnhelper_executable_issue
This monitoring issue indicates that the jspawnhelper
JVM file on the Solaris host does not have the executable permission.
To fix the Solaris JVM file permission issue, provide the executable permission to the jspawnhelper
JVM file in the JVM path using the following command:
chmod +x <jvm-path>/jspawnhelper
Where JVM path for earlier version of Java 9 is <instana-agent-dir>/jvm/jre/lib/sparcv9/jspawnhelper
and for Java 9 or later is <instana-agent-dir>/jvm/jre/lib/jspawnhelper
.