Skip to main content

Deployment Management

UniAgent Management​

On the UniAgent management page, users can view all monitored hosts, including Linux hosts and Windows hosts.

List Item Description​

The list items are described as follows:

  • Dot Color: Green indicates that the UniAgent status is normal, red indicates abnormal, blue indicates not monitored, and gray indicates no data.
  • Host Name: The name of the host.
  • Agent Type: The Agent can be deployed on either Linux or Windows systems.
  • Deployment Environment: The data center, VPC, server room, or other environment where the UniAgent host is located.
  • IP: The IP address of the host where UniAgent is located.
  • Port: The port number that UniAgent listens on.
  • CPU Cores: The number of CPU cores on the host where UniAgent is located. If UniAgent is deployed in a virtual machine or Docker, it refers to the number of CPU cores of the virtual machine or Docker itself.
  • Total Memory: The memory of the host where UniAgent is located. If UniAgent is deployed in a virtual machine or Docker, it refers to the memory of the virtual machine or Docker itself.
  • Version: UniAgent version.
  • Upgrade Status: The online upgrade status of UniAgent, including upgrading, upgrade successful, upgrade failed, etc.
  • Installation Time: UniAgent installation time.
  • Heartbeat Time: The last data update time of UniAgent, i.e., the last heartbeat time.

Operation Instructions​

Users can perform the following operations on this page:

  • View process information: Click the icon before the host name to view all process information on that host, such as process name, process PID, and communication ports. Click the PID will navigate to the process details page; click the App Name will navigate to the application analysis page; click the Instance Name will navigate to the application instance analysis page.

  • Disable UniAgent: Turn off the toggle in the Disable/Enable column, and process information will no longer be displayed under the host. The UniAgent Agent will no longer instrument restarted or newly started processes. This means it will stop collecting data on host resources, all basic component processes, and application processes on the host. Applications that have already been instrumented will continue to upload data normally.

  • Edit host: Click Edit in the Operation column to modify the host name. This name is only displayed as an alias in the Guanyun platform, and the actual host name will not change. If the full-stack data collection function was not enabled when deploying UniAgent, you can edit the host to enable it. After turning on the Full-stack Data Collection toggle, UniAgent can collect performance data for infrastructure and server. If the toggle is turned off, only infrastructure performance data will be collected (requires application restart to take effect).

  • Search host: Click the filter box at the top of the page to display the filter conditions. Select the filter condition, enter the content, and then click Query to query the host, or you can directly press the Enter key to query. It supports filtering through a single keyword, multiple keywords, a single column field, multiple values for a single column field, multiple column fields, a single tag, and multiple tags. You can click the icon before Query to view the search help. Users can set multiple filter conditions and search by Agent type, Host Name, deployment environment, IP address, and version.

  • Customize columns: Click Customize Columns in the upper right corner of the UniAgent list to select which column names you want to display. Some fields are mandatory and must be shown.

  • Only show hosts from the Last Week: By default, the UniAgent list displays information for all hosts with UniAgent installed, including those without data. Check the Only show hosts from the last week checkbox in the upper right corner of the page, and the list will only display hosts that have data in the last week. The system will filter based on the last heartbeat time of UniAgent.

  • Batch enable: Check the checkboxes in front of multiple host names, and the Batch Enable button in the upper right corner of the UniAgent list becomes active. Users may then click this button to enable UniAgents in bulk.

  • Batch disable: Check the checkboxes in front of multiple host names, and the Batch Disable button in the upper right corner of the UniAgent list becomes active. Users may then click this button to disable UniAgents in bulk. Process information will no longer be displayed under disabled hosts. The UniAgent Agent will no longer instrument restarted or newly started processes. This means it will stop collecting data on host resources, all basic component processes, and application processes on the host. Applications that have already been instrumented will continue to upload data normally.

  • Batch upgrade: Check the checkbox in front of one or more host names, and the Batch Upgrade button in the upper right corner of the UniAgent list becomes active. Users may then click the button to upgrade UniAgent in bulk. In the pop-up dialog box, select the UniAgent version to upgrade to, and then click OK to initiate the upgrade. The Upgrade Status column displays the upgrade process of UniAgent on each host. Click the status to view the upgrade log.

Upgrade Recommendation: In the same deployment environment, first perform a canary upgrade on one host. If no issues are observed, proceed with batch upgrades. It is recommended to upgrade 5–20 hosts per batch, and start the next batch only after ensuring stability.

  • Export UniAgent Data: Click Export in the upper right corner of the UniAgent list to export all UniAgent data. If filter conditions are set, the exported result is not the search result, but still all UniAgent data.

  • Add UniAgent: Click Add in the upper right corner of the UniAgent list to enter the UniAgent deployment page, where users can deploy a new UniAgent.

  • Delete UniAgent: Click Delete in the upper right corner of the UniAgent list to delete UniAgent.

Process Information​

  • Process Name: The process name is determined by the startup method and the type of process. When started with Docker, it will contain using information, with the image name after using.

    Note: Go, Python, and Node.js applications are not displayed in the process information.

    Java: There are three startup methods: non-Docker startup, Docker startup, and kubernetes startup Docker. Currently, five process types are supported: Tomcat, WebLogic, SpringBoot, Jar, and Java. Taking Tomcat as an example, as shown below:

    • Non-Docker Startup

    • Docker Startup

    • Docker + kubernetes Startup

    .NET Core: There are three startup methods: non-Docker startup, Docker startup, and kubernetes startup Docker. Since .NET Core does not require a container to start, there is only one process type. Examples of each startup method are as follows:

    • Non-Docker Startup

    • Docker Startup

    • kubernetes Startup Docker

  • PID: Process ID.

  • DockerId: If the process is started in a Docker container, the first eight digits of the DockerId are taken.

  • kubernetes pod UID: The podId of kubernetes.

  • Agent Version: Agent version.

  • App Name: Application name.

  • Port: Process listening port.

  • Instance Name: Instance name. It is divided into the following four situations (when the port cannot be obtained, the default is 0):

    • Normal Process: hostName:port.

    • Weblogic Process: ServerName:port. Since .NET Core is deployed without a container, there will be no Weblogic process, so this type of instance name will not appear.

    • Docker Container Process: The first 12 digits of ContainerId:port.

    • kubernetes Startup Docker Container Process: PodName:port.

  • Toggle: Process instrumentation disable/enable toggle. When disabled, the process will not be instrumented. The process needs to be restarted for the change to take effect.

  • Process Status

    • Pending Restart Status

      Process need restart: For processes started before UniAgent is installed, the processes must be restarted after UniAgent installation to embed the Agent.

    • Normal Instrumentation Status

    • Disable Pending Restart Status

      Process need to be restarted to remove application agent: After disabling a process, you need to restart the process on the server for the change to take effect.

    • Disabled Status

      Process is disabled: Not instrumented status.

Collector Management​

Agent Collector, abbreviated as Collector, is responsible for aggregating performance data. A Guanyun platform system can deploy multiple Agent Collectors.

Select Deployment Status in the left navigation bar, and then select the Collectors Management tab at the top to enter the Collector management page.

Collector List​

The Collector list displays Collector information with data within the last 7 days. The number after Display in the upper left corner of the list represents: Number of Collectors with data within 7 days/Total Collectors.

List Item Description​

  • IP: The IP address of the Collector.
  • Deployment Environment: The name of the data center, VPC, server room, etc.
  • Version: The version number of the Collector.
  • Protocol: The protocol version number used by the Collector.
  • Load: The CPU busyness of the host where the Collector is located. When the load value is between [0,1], it indicates that the CPU can process instructions without waiting.
  • CPU: The current CPU usage of the Collector.
  • Memory: The current memory usage and usage rate of the Collector.
  • Throughput: The average number of data collection and aggregation requests processed by the Collector per second during the selected time period, in Rps.
  • Bandwidth: The average bandwidth of data collection and aggregation requests processed by the Collector during the selected time period, in Byte/s.
  • Status: When the CPU usage is greater than or equal to 70% or the memory usage is greater than or equal to 80%, it is considered a resource-tight state. When the number of managed Agents is 0, it is considered an idle state. Other situations are normal states. If a Collector has no heartbeat for 10 minutes, it will be in an offline state.
  • Upgrade Status: The online upgrade status of the Collector, including upgrading, upgrade successful, upgrade failed, etc.
  • Start Time: The most recent start time of the Collector.
  • Heartbeat Detection Time: The most recent time when the Collector heartbeat was detected.

Operation Instructions​

Users can perform the following operations on Collectors.

  • Search Collector: Click the filter box at the top of the page to display the filter conditions. Select the filter condition, enter the content, and then click Query to query the Collector, or you can directly press the Enter key to query. It supports filtering through a single keyword, multiple keywords, a single column field, multiple values for a single column field, multiple column fields, a single tag, and multiple tags. You can click the icon before Query to view the search help. Users can set multiple filter conditions and search by deployment environment, IP address, and version.

  • View Trend Charts of Running Metrics for Each Component of the Collector: Click the icon before the host name to see the icons of each component of the Collector, including APM Collector, Infra Collector, and 12 component Agent icons.

    Click these icons to view the trend charts of CPU, memory, latency, and running time (the vertical axis represents seconds).

  • Customize columns: Click Customize Columns in the upper right corner of the list to select which column names you want to display. Some fields are mandatory and must be shown.

  • Disable or Enable Collector: Turn off the toggle in the Disable/Enable column to disable the Collector. After disabling, the Collector heartbeat is still retained for re-enabling. Connected Agents will be redistributed to other Collectors.

  • Batch Enable: Check the checkboxes in front of multiple Collectors, and the Batch Enable button in the upper right corner of the Collector list becomes available. Users can click the button to enable Collectors in bulk. After enabling, the Collectors will join the IDC (deployment environment) and share the load of Agent connections under the IDC.

  • Batch Disable: Check the checkboxes in front of multiple Collectors, and the Batch Disable button in the upper right corner of the Collector list becomes active. Users may then click the button to disable Collectors in bulk. After disabling, the Collector heartbeat is still retained for re-enabling. Connected Agents will be redistributed to other Collectors.

  • Batch Upgrade: Check the checkboxes in front of multiple Collectors, and the Batch Upgrade button in the upper right corner of the Collector list becomes active. Users may then click the button to upgrade Collectors in bulk. In the pop-up dialog box, select the Collector version to upgrade to, and then click OK to initiate the upgrade. The Upgrade Status column displays the upgrade process of the Collector. Click the status to view the upgrade log.

Upgrade Recommendation: First perform a canary upgrade on one host, if no issues are observed, and then proceed with upgrades one by one. Batch upgrades are not recommended.

  • Export Collector Data: Click Export in the upper right corner of the Collector list to export all Collector data. If filter conditions are set, the exported result is not the search result, but still all Collector data.

  • Add Collector: Click Add in the upper right corner of the Collector list to enter the Collector deployment page, where users can deploy a new Collector. For details, please refer to Collector Deployment.

  • Delete Collector: Check the Collectors to be deleted in the list, click Delete in the upper right corner of the Collector list, and all component Agent instances under the Collector, except for the host Agent, will be deleted.

Collector Host Details​

After clicking any item in the Collector list, the trend charts of CPU, various memories, bandwidth, number of managed Agents, etc. of the host where the Collector is located will be displayed below the list.

Click View Configuration in the upper right corner to view property configuration, process, and environment variable information. The specific meaning of each configuration item can be viewed in collector.properties.