.. include:: /../getting_started/vars.rst .. _section-getting_started: Getting Started Guide ********************* About the Observability Example =============================== *Observability Framework* includes a C++ example that you can use to evaluate the capabilities of this experimental product. The example is installed in your rti_workspace directory, in the ``/examples/observability/c++`` folder. This section details how the example is configured and how to run it. When you are ready to test the example, refer to the sections `Before Running the Example <#before-running-the-example>`__ and `Running the Example <#running-the-example>`__ for instructions. .. important:: *Observability Framework* is an experimental product that includes example configuration files for use with several third-party components (Prometheus, Grafana Loki, and Grafana). This release is an evaluation distribution; use it to explore the new observability features that support |Connext| applications. Do not deploy any *Observability Framework* components in production. Applications ------------ The example consists of two applications: - One application publishes simulated data generated by temperature sensors. - One application subscribes to the sensor data generated by the temperature sensors. You can run multiple publishing and subscribing applications in the same host, or in multiple hosts, within a LAN. Each publishing application can handle multiple sensors, and each subscribing application subscribes to all sensors. .. figure:: static/observability_applications.png :figwidth: 75 % :alt: Observability Applications :name: FigureObservabilityApplications :align: center To learn more about the publish/subscribe model, refer to :link_connext_dds_pro_gsg_cpp98:`Publish/Subscribe` in the *RTI Connext Getting Started Guide*. The example applications use *Monitoring Library 2.0* to forward telemetry data (logs and metrics) to *Observability Collector Service*. The collector stores this data in Prometheus (metrics) and Grafana Loki (logs) for analysis and visualization using Grafana. .. figure:: static/observability_collector_service.png :figwidth: 75 % :alt: Observability Collector Service :name: FigureObservabilityCollectorService :align: center Data Model ---------- The DDS data model for the Temperature topic used in this example is as follows: .. code-block:: text // Temperature data type struct Temperature { // ID of the sensor sending the temperature @key uint32 sensor_id; // Degrees in Celsius int32 degrees; }; Each sensor represents a different instance in the Temperature topic. For general information about data types and topics, refer to :link_connext_dds_pro_gsg_cpp11:`Introduction to DataWriters, DataReaders, and Topics` and :link_connext_dds_pro_gsg_cpp11:`Data Types` in the *RTI Connext Getting Started Guide*. DDS Entity Mapping ------------------ The Publisher application creates one DomainParticipant and n-DataWriters, where *n* is the number of sensors published by the application. This number is configurable using the command ``--sensor-count``. Each DataWriter publishes one instance. Refer to :link_connext_dds_pro_gsg_cpp11:`Keys and Instances` in the *RTI Connext Getting Started Guide* for more information on instances. The Subscriber application creates one DomainParticipant and a single DataReader to subscribe to all sensor data. Command Line Parameters ----------------------- The following command-line switches are available when starting the Publisher and Subscriber applications included in the example. Use this information as a reference when you `run the example <#running-the-example>`__. Publishing Application ^^^^^^^^^^^^^^^^^^^^^^ .. list-table:: Publishing Application :name: TablePublisherOptions :widths: 30 20 50 10 :header-rows: 1 * - Parameter - Data Type - Description - Default * - -n, --application-name - - Application name - SensorPublisher_ * - -d, --domain - - Application domain ID - 0 * - -i, --init-sensor-id - - Initial sensor ID - 0 * - -s, --sensor-count - - Sensor count. Each sensor writes one instance published by a separate DataWriter - 1 * - -o, --observability-domain - - Domain for sending telemetry data - 2 * - -c, --collector-peer - - Collector service peer - localhost * - -v, --verbosity - - How much debugging output to show, range 0-3 - 1 * - -p, --protected - N/A - Enable security - disabled The publishing applications should not publish information for the same sensor IDs. To avoid this issue, you will use the ``-i`` command-line parameter to specify the sensor ID to be used as the initial ID when `Running the Example <#running-the-example>`__. Subscribing Application ^^^^^^^^^^^^^^^^^^^^^^^ .. list-table:: Subscribing Application :name: TableSubscriberOptions :widths: 30 20 50 10 :header-rows: 1 * - Parameter - Data Type - Description - Default * - -n, --application-name - - Application name - SensorSubscriber * - -d, --domain - - Application domain ID - 0 * - -o, --observability-domain - - Domain for sending telemetry data - 2 * - -c, --collector-peer - - Collector service peer - localhost * - -v, --verbosity - - How much debugging output to show, Range 0-3 - 1 * - -p, --protected - N/A - Enable security - disabled Before Running the Example ========================== Set Up Environment Variables ---------------------------- Set up the environment variables for running and compiling the example: #. Open a command prompt window. |br| |br| #. Run this script: .. tabs:: .. group-tab:: Linux .. code-block:: console $ source /resource/scripts/rtisetenv_.bash If you're using the Z shell, run this: .. code-block:: console $ source /resource/scripts/rtisetenv_.zsh .. group-tab:: macOS .. code-block:: console $ source /resource/scripts/rtisetenv_.bash If you're using the Z shell, run this: .. code-block:: console $ source /resource/scripts/rtisetenv_.zsh .. group-tab:: Windows .. code-block:: doscon > /resource/scripts/rtisetenv_.bat ```` refers to the installation directory for |Connext|. The rtisetenv script adds the location of the SDK libraries (``/lib/``) to your library path, sets the ```` environment variable to point to ````, and puts the *RTI Code Generator* tool in your path. You may need *Code Generator* if the makefile for your architecture is not available under the *make* directory in the example. Your architecture (such as ``x64Linux3gcc7.3.0``) is the combination of processor, OS, and compiler version that you will use to build your application. For example: .. code-block:: console $ source $NDDSHOME/resource/scripts/rtisetenv_x64Linux4gcc7.3.0.bash Compile the Example ------------------- *Monitoring Library 2.0* can be used in three different ways: - **Dynamically loaded:** This method requires that the rtimonitoring2 shared library is in the library search path. - **Dynamic linking:** The application is linked with the rtimonitoring2 shared library. - **Static linking:** The application is linked with the rtimonitoring2 static library. You will compile the example using |Connext| shared libraries so that *Monitoring Library 2.0* can be dynamically loaded. The example is installed in your rti_workspace directory, in the ``/examples/observability/c++`` folder. Non-Windows Systems ^^^^^^^^^^^^^^^^^^^ To build this example on a non-Windows system, type the following in a command shell from the example directory: .. code-block:: console $ make -f make/makefile_Temperature_ DEBUG=0 If there is no makefile for your architecture in the *make* directory, you can generate it using the following ``rtiddsgen`` command: .. code-block:: console $ rtiddsgen -language C++98 -create makefiles -platform -sharedLib -sourceDir src -d ./make ./src/Temperature.idl Windows Systems ^^^^^^^^^^^^^^^ To build this example on Windows, open the appropriate solution file for your version of Microsoft Visual Studio in the win32 directory. To use dynamic linking, select **Release DLL** from the dropdown menu. .. _section-start-docker-containers: Install Observability Framework ------------------------------- Before running the example, make sure that you have installed both *Monitoring Library 2.0* and the collection, storage and visualization components. Refer to the :ref:`section-installation` section for instructions. If you want to run the example with security enabled, you must install *Observability Framework* using a secure configuration. If you did not create a secure configuration, delete the existing configuration as described in section :ref:`section-remove-observability-workspace`, then update your JSON configuration file as needed. The following sections include example configuration files you can edit for your environment. The collection, storage, and visualization components can be installed using one of two methods: - Install the components in a Linux host on the same LAN where the applications run, or - Install the components on a remote Linux host (for example, an AWS instance) reachable over the WAN using *RTI Real-Time WAN Transport*. In either of these methods, you can use a secure or non-secure configuration. To facilitate testing secure configurations where all components run on the same node (docker images, test applications, and browser), artifacts are provided in your rti_workspace directory, in the ``/examples/dds-security/`` folder. The artifacts provided to secure the https connections use the hostname "localhost". The following sections provide example JSON configurations for each of the eight configurations supported by *Observability Framework*. These configurations use the hostname "localhost", default port values, and the paths to the default security artifacts where appropriate. You can copy these examples to a local file and use them as is or customize them with your own hostname, ports, and security artifacts. For details on how to configure *Observability Framework* see section :ref:`section-configure-json-file`. .. _section-configure-operation-mode: Configure Observability Framework for the Appropriate Operation Mode ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ .. important:: The provided example configurations work only if you run ALL components (docker images, test applications, and browser) on the same host machine. If you intend to run any components (test applications or browser) on a remote machine, you must update the ``hostname`` field in the JSON configuration file to the hostname of the machine running *Observability Framework*. Additionally, if you run in secure mode, you will need to generate the https security artifacts and the DDS security artifacts as shown in :ref:`section-generating-security-artifacts`. Once you have generated your artifacts, you will need to update the ``securityConfig`` section in the JSON configuration file with the paths to these artifacts. There are eight distinct operation modes you can use to configure *Observability Framework*. These modes, described below, are based on the desired security and network environment (LAN or WAN). #. Select the operation mode for the test you want to run, then edit your JSON configuration file with the selected content. For example, if you want to test on a LAN without security, copy the example JSON fom section :ref:`section-non-secure-lan-example` to the ``config.json`` file. |br| |br| #. If desired, modify the hostname, ports, or security artifact paths in the ``config.json`` file. For example, to use port 9091 for Prometheus, change the "prometheusPort" field in the ``config.json`` file from 9090 to 9091. |br| |br| #. Run the rtiobservability script to apply your *Observability Framework* configuration. .. tabs:: .. group-tab:: Linux .. code-block:: console $ rtiobservability -c config.json If you have already configured *Observability Framework* in a different operation mode than the one you want to test, you must first remove the existing workspace as described in section :ref:`section-remove-observability-workspace`. Example LAN configurations """""""""""""""""""""""""" :numref:`TableDockerContainerLanOperationModes` lists the four LAN configurations supported by *Observability Framework*. .. list-table:: Docker Container LAN Operation Modes :name: TableDockerContainerLanOperationModes :widths: 30 20 50 10 :header-rows: 1 * - Configuration Name - Network - Data Storage - Security * - NonSecureLAN - LAN - Prometheus and Grafana Loki - No * - SecureLAN - LAN - Prometheus and Grafana Loki - Yes * - NonSecureOTelLAN - LAN - Multiple through OpenTelemetry Collector - No * - SecureOTelLAN - LAN - Multiple through OpenTelemetry Collector - Yes .. _section-non-secure-lan-example: Non-Secure LAN Configuration ++++++++++++++++++++++++++++ This example configures *Observability Framework* with hostname "localhost", default ports, and security disabled. *Observability Collector Service* will use a LAN configuration. .. code-block:: JSON { "hostname": "localhost", "observabilityDomain": 2, "lgpStackConfig": { "grafanaPort": 3000, "prometheusPort": 9090, "lokiPort": 3100 }, "collectorConfig": { "prometheusExporterPort": 19090, "controlPort": 19098 } } .. _section-secure-lan-example: Secure LAN Configuration ++++++++++++++++++++++++ This example configures *Observability Framework* with hostname "localhost", default ports, and security enabled. *Observability Collector Service* will use a LAN configuration. .. code-block:: JSON { "hostname": "localhost", "observabilityDomain": 2, "lgpStackConfig": { "grafanaPort": 3000, "prometheusPort": 9090, "lokiPort": 3100 }, "collectorConfig": { "prometheusExporterPort": 19090, "controlPort": 19098 }, "securityConfig": { "basicAuthUsername": "user", "basicAuthPassword": "userpassword", "httpsSecurity": { "caCertificate": "/examples/dds_security/cert/ecdsa01/ca/ecdsa01RootCaCert.pem", "serverCertificate": "/examples/dds_security/cert/ecdsa01/https/ecdsa01Https01Cert.pem", "serverKey": "/examples/dds_security/cert/ecdsa01/https/ecdsa01Https01Key.pem" }, "ddsSecurity": { "identityCaCertificate": "/examples/dds_security/cert/ecdsa01/ca/ecdsa01RootCaCert.pem", "permissionsCaCertificate": "/examples/dds_security/cert/ecdsa01/ca/ecdsa01RootCaCert.pem", "identityCertificate": "/examples/dds_security/cert/ecdsa01/identities/ecdsa01Peer01Cert.pem", "identityKey": "/examples/dds_security/cert/ecdsa01/identities/ecdsa01Peer01Key.pem", "signedPermissionsFile": "/examples/dds_security/xml/signed/signed_ObservabilityCollectorServicePermissions.p7s", "signedGovernanceFile": "/examples/dds_security/xml/signed/signed_ObservabilityGovernance.p7s" } } } .. _section-non-secure-otel-lan-example: Non-Secure OTel LAN Configuration +++++++++++++++++++++++++++++++++ This example configures *Observability Framework* with hostname "localhost", default ports, and security disabled. *Observability Collector Service* will use a LAN configuration and the OpenTelemetry exporter. The OpenTelemetry Collector routes telemetry data from the *Observability Collector Service* OpenTelemetry exporter to Prometheus and Loki. .. code-block:: JSON { "hostname": "localhost", "observabilityDomain": 2, "lgpStackConfig": { "grafanaPort": 3000, "prometheusPort": 9090, "lokiPort": 3100 }, "collectorConfig": { "prometheusExporterPort": 19090, "controlPort": 19098 }, "otelConfig": { "otelHttpReceiverPort": 4318 } } .. _section-secure-otel-lan-example: Secure OTel LAN Configuration +++++++++++++++++++++++++++++ This example configures *Observability Framework* with hostname "localhost", default ports, and security enabled. *Observability Collector Service* will use a LAN configuration and the OpenTelemetry exporter. The OpenTelemetry Collector routes telemetry data from the *Observability Collector Service* OpenTelemetry exporter to Prometheus and Loki. .. code-block:: JSON { "hostname": "localhost", "observabilityDomain": 2, "lgpStackConfig": { "grafanaPort": 3000, "prometheusPort": 9090, "lokiPort": 3100 }, "collectorConfig": { "prometheusExporterPort": 19090, "controlPort": 19098 }, "otelConfig": { "otelHttpReceiverPort": 4318 }, "securityConfig": { "basicAuthUsername": "user", "basicAuthPassword": "userpassword", "httpsSecurity": { "caCertificate": "/examples/dds_security/cert/ecdsa01/ca/ecdsa01RootCaCert.pem", "serverCertificate": "/examples/dds_security/cert/ecdsa01/https/ecdsa01Https01Cert.pem", "serverKey": "/examples/dds_security/cert/ecdsa01/https/ecdsa01Https01Key.pem" }, "ddsSecurity": { "identityCaCertificate": "/examples/dds_security/cert/ecdsa01/ca/ecdsa01RootCaCert.pem", "permissionsCaCertificate": "/examples/dds_security/cert/ecdsa01/ca/ecdsa01RootCaCert.pem", "identityCertificate": "/examples/dds_security/cert/ecdsa01/identities/ecdsa01Peer01Cert.pem", "identityKey": "/examples/dds_security/cert/ecdsa01/identities/ecdsa01Peer01Key.pem", "signedPermissionsFile": "/examples/dds_security/xml/signed/signed_ObservabilityCollectorServicePermissions.p7s", "signedGovernanceFile": "/examples/dds_security/xml/signed/signed_ObservabilityGovernance.p7s" } } } Example WAN configurations """""""""""""""""""""""""" :numref:`TableDockerContainerWANOperationModes` lists the four WAN configurations supported by *Observability Framework*. .. list-table:: Docker Container WAN Operation Modes :name: TableDockerContainerWANOperationModes :widths: 30 20 50 10 :header-rows: 1 * - Configuration Name - Network - Data Storage - Security * - NonSecureWAN - WAN - Prometheus and Grafana Loki - No * - SecureWAN - WAN - Prometheus and Grafana Loki - Yes * - NonSecureOTelWAN - WAN - Multiple through OpenTelemetry Collector - No * - SecureOTelWAN - WAN - Multiple through OpenTelemetry Collector - Yes .. _section-non-secure-wan-example: Non-Secure WAN Configuration ++++++++++++++++++++++++++++ This example configures *Observability Framework* with hostname "localhost", default ports, and security disabled. *Observability Collector Service* will use a WAN configuration with port 30000. .. code-block:: JSON { "hostname": "localhost", "observabilityDomain": 2, "lgpStackConfig": { "grafanaPort": 3000, "prometheusPort": 9090, "lokiPort": 3100 }, "collectorConfig": { "prometheusExporterPort": 19090, "controlPort": 19098, "rtwPort": 30000 } } .. _section-secure-wan-example: Secure WAN Configuration ++++++++++++++++++++++++ This example configures *Observability Framework* with hostname "localhost", default ports, and security enabled *Observability Collector Service* will use a WAN configuration with port 30000. .. code-block:: JSON { "hostname": "localhost", "observabilityDomain": 2, "lgpStackConfig": { "grafanaPort": 3000, "prometheusPort": 9090, "lokiPort": 3100 }, "collectorConfig": { "prometheusExporterPort": 19090, "controlPort": 19098, "rtwPort": 30000 }, "securityConfig": { "basicAuthUsername": "user", "basicAuthPassword": "userpassword", "httpsSecurity": { "caCertificate": "/examples/dds_security/cert/ecdsa01/ca/ecdsa01RootCaCert.pem", "serverCertificate": "/examples/dds_security/cert/ecdsa01/https/ecdsa01Https01Cert.pem", "serverKey": "/examples/dds_security/cert/ecdsa01/https/ecdsa01Https01Key.pem" }, "ddsSecurity": { "identityCaCertificate": "/examples/dds_security/cert/ecdsa01/ca/ecdsa01RootCaCert.pem", "permissionsCaCertificate": "/examples/dds_security/cert/ecdsa01/ca/ecdsa01RootCaCert.pem", "identityCertificate": "/examples/dds_security/cert/ecdsa01/identities/ecdsa01Peer01Cert.pem", "identityKey": "/examples/dds_security/cert/ecdsa01/identities/ecdsa01Peer01Key.pem", "signedPermissionsFile": "/examples/dds_security/xml/signed/signed_ObservabilityCollectorServicePermissions.p7s", "signedGovernanceFile": "/examples/dds_security/xml/signed/signed_ObservabilityGovernance.p7s" } } } .. _section-non-secure-otel-wan-example: Non-Secure OTel WAN Configuration +++++++++++++++++++++++++++++++++ This example configures *Observability Framework* with hostname "localhost", default ports, and security disabled. *Observability Collector Service* will use a WAN configuration with port 30000 and the OpenTelemetry exporter. The OpenTelemetry Collector routes telemetry data from *Observability Collector Service* OpenTelemetry exporter to Prometheus and Loki. .. code-block:: JSON { "hostname": "localhost", "observabilityDomain": 2, "lgpStackConfig": { "grafanaPort": 3000, "prometheusPort": 9090, "lokiPort": 3100 }, "collectorConfig": { "prometheusExporterPort": 19090, "controlPort": 19098, "rtwPort": 30000 }, "otelConfig": { "otelHttpReceiverPort": 4318 } } .. _section-secure-otel-wan-example: Secure OTel WAN Configuration +++++++++++++++++++++++++++++ This example configures *Observability Framework* with hostname "localhost", default ports, and security enabled. *Observability Collector Service* will use a WAN configuration with port 30000 and the OpenTelemetry exporter. The OpenTelemetry Collector routes telemetry data from the *Observability Collector Service* OpenTelemetry exporter to Prometheus and Loki. .. code-block:: JSON { "hostname": "localhost", "observabilityDomain": 2, "lgpStackConfig": { "grafanaPort": 3000, "prometheusPort": 9090, "lokiPort": 3100 }, "collectorConfig": { "prometheusExporterPort": 19090, "controlPort": 19098, "rtwPort": 30000 }, "otelConfig": { "otelHttpReceiverPort": 4318 }, "securityConfig": { "basicAuthUsername": "user", "basicAuthPassword": "userpassword", "httpsSecurity": { "caCertificate": "/examples/dds_security/cert/ecdsa01/ca/ecdsa01RootCaCert.pem", "serverCertificate": "/examples/dds_security/cert/ecdsa01/https/ecdsa01Https01Cert.pem", "serverKey": "/examples/dds_security/cert/ecdsa01/https/ecdsa01Https01Key.pem" }, "ddsSecurity": { "identityCaCertificate": "/examples/dds_security/cert/ecdsa01/ca/ecdsa01RootCaCert.pem", "permissionsCaCertificate": "/examples/dds_security/cert/ecdsa01/ca/ecdsa01RootCaCert.pem", "identityCertificate": "/examples/dds_security/cert/ecdsa01/identities/ecdsa01Peer01Cert.pem", "identityKey": "/examples/dds_security/cert/ecdsa01/identities/ecdsa01Peer01Key.pem", "signedPermissionsFile": "/examples/dds_security/xml/signed/signed_ObservabilityCollectorServicePermissions.p7s", "signedGovernanceFile": "/examples/dds_security/xml/signed/signed_ObservabilityGovernance.p7s" } } } Start the Collection, Storage, and Visualization Docker Containers ------------------------------------------------------------------ The Docker containers used for data collection, storage, and visualization can either be run in a Linux host on the same LAN where the applications run or they can be installed on a remote Linux host (for example, an AWS instance) reachable over the WAN using *RTI Real-Time WAN Transport*. There may be different licensing requirements depending on the configuration (LAN/WAN, Secure/Non-Secure) you have chosen to run. For details on the license requirements and instructions on how to run the containers, see section :ref:`section-initializing-and-running-docker`. Running the Example =================== Before running the example, select the appropriate values for the following deployment command-line options as needed: .. list-table:: Optional Command-Line Parameters :name: TableCommandLineOptions :widths: 30 50 20 :header-rows: 1 * - Parameter - Description - Default Value * - ``--observability-domain`` - Use this command-line option if you want to overwrite the default domain ID used by *Monitoring Library 2.0* to send telemetry data to *Observability Collector Service*. - 2 * - ``--collector-peer`` - If you run *Observability Collector Service* in a different host from the applications, use this command-line option to provide the address of the service. For example, 192.168.1.1 (for LAN), or udpv4_wan://10.56.78.89:16000 (for WAN). - localhost In addition, if you run the applications in different hosts and multicast is not available, use the NDDS_DISCOVERY_PEERS environment to configure the peers where the applications run. For simplicity, the instructions in this section assume that you are running the applications and the Docker containers used by *Observability Framework* on the same host with the default observability domain. .. _section-start-the-applications: Start the Applications ---------------------- This example assumes ``x64Linux4gcc7.3.0`` as the architecture. The following steps include instructions for non-secure and secure tests. #. In a new browser window, go to http[s]://localhost:3000 and log in using your Grafana dashboard credentials. Note the use of ``https`` if you are running a secure configuration. The default Grafana dashboard credentials are ``admin:admin`` for non-secure configurations and ``user:userpassword`` for secure configurations. .. figure:: static/grafana_dashboard.png :figwidth: 100 % :alt: Grafana dashboard :name: FigureGrafanaDashboard :align: center Note that at this point, no DDS applications are running. #. From the example directory, open two terminals and start two instances of the application that publishes temperature sensor data. The command and resulting output for each instance are shown below. The ``-i`` parameter specifies the sensor ID that will be used. The ``-n`` parameter assigns a name to the application. This name will be used when sending the commands in :ref:`section-change-verbosity`. Note the addition of the ``-p`` parameter to enable security for secure configurations. The first instance creates two sensors. .. tabs:: .. group-tab:: Non-Secure LAN .. code-block:: console $ ./objs/x64Linux4gcc7.3.0/Temperature_publisher -n SensorPublisher_1 -d 57 -i 0 -s 2 -v 2 ********************************************************** ******** Temperature Sensor Publisher App **************** ********************************************************** Running with parameters: Application Resource Name: /applications/SensorPublisher_1 Domain ID: 57 Init Sensor ID: 0 Sensor Count: 2 Observability Domain: 2 Collector Peer: udpv4://localhost Verbosity: 2 Security: false Running with QOS: Temperature_Profile_With_Monitoring2_Over_LAN Command> .. group-tab:: Secure LAN .. code-block:: console $ ./objs/x64Linux4gcc7.3.0/Temperature_publisher -n SensorPublisher_1 -d 57 -i 0 -s 2 -p -v 2 ********************************************************** ******** Temperature Sensor Publisher App **************** ********************************************************** Running with parameters: Application Resource Name: /applications/SensorPublisher_1 Domain ID: 57 Init Sensor ID: 0 Sensor Count: 2 Observability Domain: 2 Collector Peer: udpv4://localhost Verbosity: 2 Security: true Running with QOS: Secure_Temperature_Profile_With_Monitoring2_Over_LAN Command> The second instance creates one sensor. .. tabs:: .. group-tab:: Non-Secure LAN .. code-block:: console $ ./objs/x64Linux4gcc7.3.0/Temperature_publisher -n SensorPublisher_2 -d 57 -i 2 -s 1 -v 2 ********************************************************** ******** Temperature Sensor Publisher App **************** ********************************************************** Running with parameters: Application Resource Name: /applications/SensorPublisher_2 Domain ID: 57 Init Sensor ID: 2 Sensor Count: 1 Observability Domain: 2 Collector Peer: udpv4://localhost Verbosity: 2 Security: false Running with QOS: Temperature_Profile_With_Monitoring2_Over_LAN Command> .. group-tab:: Secure LAN .. code-block:: console $ ./objs/x64Linux4gcc7.3.0/Temperature_publisher -n SensorPublisher_2 -d 57 -i 2 -s 1 -p -v 2 ********************************************************** ******** Temperature Sensor Publisher App **************** ********************************************************** Running with parameters: Application Resource Name: /applications/SensorPublisher_1 Domain ID: 57 Init Sensor ID: 2 Sensor Count: 1 Observability Domain: 2 Collector Peer: udpv4://localhost Verbosity: 2 Security: true Running with QOS: Secure_Temperature_Profile_With_Monitoring2_Over_LAN Command> #. From the example directory, open a new terminal and start one instance of the application that subscribes to temperature sensor data. .. tabs:: .. group-tab:: Non-Secure LAN .. code-block:: console $ ./objs/x64Linux4gcc7.3.0/Temperature_subscriber -n SensorSubscriber -d 57 -v 2 ********************************************************** ******** Temperature Sensor Subscriber App **************** ********************************************************** Running with parameters: Application Resource Name: /applications/SensorSubscriber Domain ID: 57 Observability Domain: 2 Collector Peer: udpv4://localhost Verbosity: 2 Security: false Running with QOS: Temperature_Profile_With_Monitoring2_Over_LAN Command> .. group-tab:: Secure LAN .. code-block:: console $ ./objs/x64Linux4gcc7.3.0/Temperature_subscriber -n SensorSubscriber -d 57 -p -v 2 ********************************************************** ******** Temperature Sensor Subscriber App **************** ********************************************************** Running with parameters: Application Resource Name: /applications/SensorSubscriber Domain ID: 57 Observability Domain: 2 Collector Peer: udpv4://localhost Verbosity: 2 Security: true Running with QOS: Secure_Temperature_Profile_With_Monitoring2_Over_LAN Command> .. note:: The two Publisher applications and the Subscriber application are started with verbosity set to WARNING (-v 2). You may see any of the following warnings on the console output. These warnings are expected. .. code-block:: console WARNING [0x01017774,0xFF40EEF6,0xEC566CA8:0x000001C1{Domain=2}|ENABLE|LC:DISC]NDDS_Transport_UDPv4_Socket_bind_with_ip:0X1EE6 in use WARNING [0x01017774,0xFF40EEF6,0xEC566CA8:0x000001C1{Domain=2}|ENABLE|LC:DISC]NDDS_Transport_UDPv4_SocketFactory_create_receive_socket:invalid port 7910 WARNING [0x01017774,0xFF40EEF6,0xEC566CA8:0x000001C1{Domain=2}|ENABLE|LC:DISC]NDDS_Transport_UDP_create_recvresource_rrEA:!create socket WARNING [0x010175D0,0x7A41F985,0xF3813392:0x000001C1{Name=Temperature DomainParticipant,Domain=57}|ENABLE|LC:DISC]NDDS_Transport_UDPv4_Socket_bind_with_ip:0X549C in use WARNING [0x010175D0,0x7A41F985,0xF3813392:0x000001C1{Name=Temperature DomainParticipant,Domain=57}|ENABLE|LC:DISC]NDDS_Transport_UDPv4_SocketFactory_create_receive_socket:invalid port 21660 WARNING [0x010175D0,0x7A41F985,0xF3813392:0x000001C1{Name=Temperature DomainParticipant,Domain=57}|ENABLE|LC:DISC]NDDS_Transport_UDP_create_recvresource_rrEA:!create socket WARNING [0x010175D0,0x7A41F985,0xF3813392:0x000001C1{Name=Temperature DomainParticipant,Domain=57}|ENABLE|LC:DISC]DDS_DomainParticipantDiscovery_add_peer:no peer locators for: peer descriptor(s) = "builtin.shmem://", transports = "", enabled_transports = "" Your Grafana dashboard should now provide information about the new Hosts, Applications, and DDS entities (Participants, DataWriters, and DataReaders). There should be 1 Host, 3 Applications, 3 Participants, 3 DataWriters, and 1 DataReader. .. figure:: static/grafana_dashboard_active.png :figwidth: 100 % :alt: Active Grafana dashboard :name: FigureGrafanaDashboardActive :align: center The Grafana main dashboard pictured above indicates that the system is healthy. You may sees warnings in the log section related to the reservation of communication ports. These warnings are expected. You can select the Warnings panel to visualize them. Next, you will introduce different failures that will affect the system's health. Changing the Time Range in Dashboards ------------------------------------- While running the examples, you can change the time range in the dashboards to reduce or expand the amount of history data displayed. Use the time picker dropdown at the top right to change the time range in any dashboard. .. figure:: static/grafana_time_range_select.png :figwidth: 100 % :alt: Grafana time range selector :name: FigureDashboardTimeRangeSelect :align: center The time picker includes a predefined list of time ranges to choose from. If you want to use a custom time range, enter the desired range in the **From** field. Use the format "now-< custom time >," where < custom time > is a unit of time; Grafana supports m-minute, h-hour, and d-day time units. For example, to show a custom range of one minute, enter "now-1m" in the **From** field, then select **Apply Time Range**. .. figure:: static/grafana_time_range_select_2.png :figwidth: 100 % :alt: Grafana time range selector detail :name: FigureDashboardTimeRangeSelectDetail :align: center .. .. note:: The time range may be changed on any dashboard, but all changes are temporary and will reset to 1 hour when you return to the **Alert Home** dashboard. Changes to the time range made in the **Alert Home** dashboard are unique in that the selected time range will be propagated to other dashboards as you navigate through the hierarchy. Simulate Sensor Failure ----------------------- The DataWriters in each application are expected to send sensor data every second, and the DataReader expects to receive sensor data from each sensor every second. This QoS contract is enforced with the Deadline QoS Policy set in USER_QOS_PROFILES.xml. Refer to :link_connext_dds_pro_gsg_cpp11:`Deadline QoS Policy` in the *RTI Connext Getting Started Guide* for basic information, or :link_connext_dds_pro_um:`DEADLINE QoSPolicy<#users_manual/DEADLINE_QosPolicy.htm>` in the *RTI Connext Core User's Manual* for detailed information. .. code-block:: text 1 0 To simulate a failure in the sensor with ID 0, enter the following command in the first Temperature_publisher instance: .. code-block:: text Command> stop 0 The Grafana dashboard updates to indicate the sensor failure. The dashboard does not update immediately; you may have to wait a few seconds to see the change reflecting the sensor failure as a Delay error. That error is expected because the deadline policy was violated when you stopped the sensor with ID 0. .. figure:: static/dashboard_delays_error.png :figwidth: 100 % :alt: Grafana dashboard with delays error :name: FigureDashboardDelays :align: center The Grafana dashboards are hierarchical. Now that you know something is happening related to latency (or delays), you can get additional details to determine the root cause. Select the **Delays** panel to drill down to the next level and get more information about the error. The second level of the Grafana dashboard indicates that there were deadline errors, which were generated by both the DataWriters associated with the sensors and the DataReader expecting sensor data at a specific rate. Still, we do not know which sensor the problem originated from. To determine that, we have to go one level deeper; select the **Missed Writer Deadline** panel to see which DataWriter caused the problem. .. figure:: static/dashboard_delays_level_2.png :figwidth: 100 % :alt: Grafana dashboard delays level 2 :name: FigureDashboardDelays2 :align: center The third level of the Grafana dashboard provides a list of entities providing deadline information. In this case we see three entities, or DataWriters, each associated with a sensor. We also see the first entity is failing. But what sensor does that entity represent? Looking at the DataWriter Name column, we can see that the failing sensor has the name “Sensor with ID=0”. This name is set using the :link_connext_dds_pro_um:`EntityName QoS Policy<#users_manual/ENTITY_NAME_QosPolicy__DDS_Extension_.htm>` when creating a DataWriter. If you want additional information, such as the machine where the sensor DataWriter is located, select the **Sensor with ID-0** link in the DataWriter Name column. .. figure:: static/dashboard_delays_level_3.png :figwidth: 100 % :alt: Grafana dashboard delays level 3 :name: FigureDashboardDelays3 :align: center The fourth and last level of the Grafana dashboard provides detailed information about an individual entity, including location-related information such as Host Name and Process Id. .. figure:: static/dashboard_delays_level_4.png :figwidth: 100 % :alt: Grafana dashboard delays level 4 :name: FigureDashboardDelays4 :align: center In addition, this last level provides information about individual metrics for the entity. Scroll down to see the details of the missed deadline metric. .. figure:: static/dashboard_delays_metrics.png :figwidth: 100 % :alt: Grafana dashboard delays metrics :name: FigureDashboardMetrics :align: center Next, restore the health of the failing sensor to review how the dashboard changes. Restart the first Temperature_publisher instance using the command ``start 0``. .. code-block:: text Command> start 0 Go back to the **Alert Home** dashboard to confirm that the sensor becomes healthy. After a few seconds, the **Delays** panel should indicate the sensor is healthy. However, part of the Delay panel is still red. Depending on the selected dashboard time range, the sensor was likely unhealthy for part of the time displayed.. .. figure:: static/dashboard_healthy_sensor.png :figwidth: 100 % :alt: Grafana dashboard healthy sensor :name: FigureDashboardHealthySensor :align: center Simulate Slow Sensor Data Consumption ------------------------------------- A subscribing application can be configured to consume sensor data at a lower rate than the publication rate. In a real scenario, this could occur if the subscribing application becomes CPU bound. This scenario simulates a problem with the subscribing application; a bug in the application logic makes it slow in processing sensor data. To test this failure, enter the ``slow_down`` command in the Temperature_subscriber instance: .. code-block:: text Command> slow_down After some seconds, the Grafana dashboard displays two new system errors related to saturation and unexpected data losses. Because the DataReader is not able to keep up with the sensor data, the dashboard indicates that there are potential data losses. At the same time, being unable to keep up with the sensor data could be a saturation sign. For example, the DataReader may consume 100% of the CPU due to the application bug. .. figure:: static/dashboard_slow_errors.png :figwidth: 100 % :alt: Grafana dashboard slow sensor errors :name: FigureDashboardSlow :align: center As you did when testing the sensor failure, select the displayed errors to navigate the dashboard hierarchy and determine the root cause of the problem. To go to the second level, select the **Data Loss** panel to see the reason for the losses. Because you slowed the subscriber application, the DataReader is not able to read fast enough. The **Dropped Samples due to History Depth** metric reveals the type of failure. Select the red errors to drill down and review further details about the problem. .. figure:: static/dashboard_slow_level_2.png :figwidth: 100 % :alt: Grafana dashboard slow sensor error level 2 :name: FigureDashboardSlowLevel2 :align: center After reviewing the errors, restore the health of the failing DataReader. In the Temperature_subscriber application, enter the ``speed_up`` command. .. code-block:: text Command> speed_up In Grafana, go back to the home dashboard and wait until the system becomes healthy again. After a few seconds, the Saturation and Data Loss panels should indicate a healthy system. Also, adjust the time window to one minute and wait until all the system status panels are green again. .. figure:: static/dashboard_slow_healthy.png :figwidth: 100 % :alt: Grafana dashboard slow sensor healthy :name: FigureDashboardSlowHealthy :align: center Simulate Time Synchronization Failures -------------------------------------- The DataReaders created by the subscribing applications expect that clocks are synchronized. The source timestamp associated with a sensor sample by the Publisher should not be farther in the future from the reception timestamp than a configurable tolerance. This behavior is configured using the DestinationOrder QoS Policy set in USER_QOS_PROFILES.xml. .. code-block:: text BY_SOURCE_TIMESTAMP_DESTINATIONORDER_QOS 1 0 This final simulation demonstrates how to use logging information to troubleshoot problems. In this scenario, you'll create a clock synchronization issue in the first instance of Temperature_publisher. The clock will move forward in the future by a few seconds, causing the DataReader to drop some sensor samples from the publishing application. To simulate this scenario, enter ``clock_forward 0`` in the first Temperature_publisher instance. .. code-block:: text Command> clock_forward 0 After some seconds, three panels in the system status section will turn red: **Data Loss**, **System Errors**, and **Delays**. Each is affected by the same underlying problem. You can select the red errors to drill down through the dashboard hierarchy and determine the root cause of the problem. First, select the **Data Loss** panel to see the reason for the error. The DataReader dropped samples coming from one or more of the DataWriters due to time synchronization issues. .. figure:: static/dashboard_time_sync_error.png :figwidth: 100 % :alt: Grafana dashboard time sync error :name: FigureDashboardTimeSyncError :align: center This error indicates that the DataReader in the subscribing application dropped some samples, but can't yet identify the problem sensor or DataWriter. To determine that, select the **Dropped Samples due to Time Sync** panel. .. figure:: static/dashboard_time_sync_level_2.png :figwidth: 100 % :alt: Grafana dashboard time sync error level 2 :name: FigureDashboardTimeSyncErrorLevel2 :align: center At this level, you can locate the DataReader reporting the error, but not the DataWriter causing it. Select the **TemperatureSensor** link in the DataReader Name column to go one more level down. .. figure:: static/dashboard_time_sync_level_3.png :figwidth: 100 % :alt: Grafana dashboard time sync error level 3 :name: FigureDashboardTimeSyncErrorLevel3 :align: center On the endpoint dashboard, there is one log warning associated with the DataReader reporting time synchronization issues. Select the red **Log Warning** to view the warning message logged by the DataReader. .. figure:: static/dashboard_time_sync_level_4.png :figwidth: 100 % :alt: Grafana dashboard time sync error level 4 :name: FigureDashboardTimeSyncErrorLevel4 :align: center This warning message provides information about the GUID of the DataWriter that published the sensor data that was dropped due to time synchronization issues. But how do we locate the DataWriter from its GUID? .. figure:: static/dashboard_time_sync_warning.png :figwidth: 100 % :alt: Grafana dashboard time sync error warning :name: FigureDashboardTimeSyncErrorWarning :align: center Note the highlighted RECEIVE FROM GUID in the log message. This represents the corresponding DataWriter that created the warning. (You can copy this GUID at this point). .. figure:: static/dashboard_time_sync_warning_highlighted.png :figwidth: 100 % :alt: Grafana dashboard time sync error warning highlights :name: FigureDashboardTimeSyncErrorWarningHighlighted :align: center Select the **DataWriters** panel to view a list of the running DataWriters. Now that we have a list of DataWriters, we can compare their GUIDs with the GUID in the log message to find the problem DataWriter. In this case the list does not have a lot of entries, so you can search manually. .. figure:: static/dashboard_datawriters.png :figwidth: 100 % :alt: Grafana dashboard datawriters list :name: FigureDashboardDatawriters :align: center However, when the number of entries is large, you can click on the funnel icon next to the **GUID** label to filter the list to the one writer with time synchronization issues by typing in the GUID or pasting the value copied from the log message. .. figure:: static/dashboard_datawriter_guid.png :figwidth: 100 % :alt: Grafana dashboard datawriter GUID :name: FigureDashboardDatawriterGUID :align: center Finally, select the problem DataWriter to learn its identity. .. figure:: static/dashboard_datawriter_identity.png :figwidth: 100 % :alt: Grafana dashboard datawriter identity :name: FigureDashboardDatawriterIdentity :align: center The problem DataWriter corresponds to sensor 0. You have successfully done root cause analysis by correlating metrics and logging. .. _section-change-verbosity: Change the Application Logging Verbosity ----------------------------------------------------------------------------- *Monitoring Library 2.0* has two verbosity settings. * **Collection verbosity** controls the level of log messages an application generates. * **Forwarding verbosity** controls the level of log messages an application forwards to the *Observability Collector Service* (making the messages visible in the dashboard). For additional information on logging, refer to :ref:`section-telemetry-logs`. By default, *Monitoring Library 2.0* only forwards error and warning log messages, even if the applications generate more verbose logging. Forwarding messages at a higher verbosity for all applications may saturate the network and the different *Observability Framework* components, such as *Observability Collector Service* and the logging aggregation backend (Grafana Loki in this example). However, in some cases you may want to change the logging Collection verbosity and/or the Forwarding verbosity for specific applications to obtain additional information when doing root cause analysis. In this section you will increase both the Collection and Forwarding verbosity levels for the first publishing application using a remote command. To do that, you will use the application resource name generated by using the ``-n`` command-line option. The three applications have the following names: * /applications/SensorPublisher_1 * /applications/SensorPublisher_2 * /applications/SensorSubscriber To change the Collection verbosity: #. From the Alert Home dashboard, select the **Applications** indicator to open the Application List dashboard. .. figure:: static/dashboard_entity_counts_bar.png :figwidth: 100 % :alt: Grafana dashboard entity counts bar :name: FigureDashboardEntityCountsBar :align: center #. From the Application List dashboard, select the **SensorPublisher_1** link to open the Alert Application Status dashboard. .. figure:: static/dashboard_application_list.png :figwidth: 100 % :alt: Grafana dashboard application list :name: FigureDashboardApplicationList :align: center #. From the Alert Application Status dashboard, select the Middleware Collection log verbosity to open the Log Control dashboard. .. figure:: static/dashboard_log_control_links.png :figwidth: 100 % :alt: Grafana dashboard log control links :name: FigureDashboardLogControlLinks :align: center #. From the Log Control dashboard, select **DEBUG** from the **Log Collection Verbosity** drop down menu. .. figure:: static/dashboard_log_collection_dropdown.png :figwidth: 100 % :alt: Grafana dashboard log collection dropdown :name: FigureDashboardLogCollectionDropdown :align: center Note that the verbosity setting color changes to yellow, indicating a change has been made. Also, the Set Collection Verbosity button is blue and enabled. .. figure:: static/dashboard_log_collection_debug_setting.png :figwidth: 100 % :alt: Grafana dashboard log collection debug setting :name: FigureDashboardLogCollectionDebugSetting :align: center #. Select the **Set Collection Verbosity** button. When prompted to confirm the change, select **Confirm** to set the Collection Verbosity level to DEBUG at the application. .. figure:: static/dashboard_log_control_confirmation_dialog.png :figwidth: 100 % :alt: Grafana dashboard log control confirmation dialog :name: FigureDashboardLogControlConfirmationDialog :align: center The application Collection verbosity is now DEBUG. If you examine the terminal window for SensorPublisher_1, you will see something similar to the following. .. code-block:: console DEBUG MIGInterpreter_parse:ACK from 0X1013131,0X4856CAC6 COMMENDSrWriterService_onSubmessage:[1689963305,136481175] writer oid 0x80002102 receives ACKNACK from reader 0x1013131.4856cac6.bd3f33f5.80000007 for lead [(0000000000,00013981)] bitcount(0), epoch(41048), isPureNack(0) DEBUG COMMENDActiveFacadeReceiver_loop:rCoTemnt##02Rcv returning message loan DEBUG NDDS_Transport_UDP_receive_rEA:rCoTemnt##02Rcv blocking on 0X549D DEBUG NDDS_Transport_UDP_receive_rEA:rCoTemnt##02Rcv received 64 bytes from 0X100007F|40284 DEBUG RTINetioReceiver_receiveFast:rCoTemnt##02Rcv received 64 bytes DEBUG COMMENDActiveFacadeReceiver_loop:rCoTemnt##02Rcv parsing message DEBUG MIGInterpreter_parse:INFO_DST from 0X1013131,0X4856CAC6 DEBUG MIGInterpreter_parse:ACK from 0X1013131,0X4856CAC6 DEBUG COMMENDActiveFacadeReceiver_loop:rCoTemnt##02Rcv returning message loan DEBUG NDDS_Transport_UDP_receive_rEA:rCoTemnt##02Rcv blocking on 0X549D DEBUG NDDS_Transport_UDP_receive_rEA:rCoTemnt##02Rcv received 64 bytes from 0X100007F|40284 DEBUG RTINetioReceiver_receiveFast:rCoTemnt##02Rcv received 64 bytes DEBUG COMMENDActiveFacadeReceiver_loop:rCoTemnt##02Rcv parsing message DEBUG MIGInterpreter_parse:INFO_DST from 0X1013131,0X4856CAC6 DEBUG MIGInterpreter_parse:ACK from 0X1013131,0X4856CAC6 DEBUG RTIEventActiveGeneratorThread_loop:rCoTemnt####Evt gathering events DEBUG RTIEventActiveGeneratorThread_loop:rCoTemnt####Evt firing events DEBUG COMMENDActiveFacadeReceiver_loop:rCoTemnt##02Rcv returning message loan DEBUG NDDS_Transport_UDP_receive_rEA:rCoTemnt##02Rcv blocking on 0X549D DEBUG RTIEventActiveGeneratorThread_loop:rCoTemnt####Evt rescheduling events DEBUG RTIEventActiveGeneratorThread_loop:rCoTemnt####Evt sleeping {00000000,1B5E7420} DEBUG NDDS_Transport_UDP_receive_rEA:rCoObsnt##00Rcv received 292 bytes from 0X100007F|46993 DEBUG RTINetioReceiver_receiveFast:rCoObsnt##00Rcv received 292 bytes DEBUG COMMENDActiveFacadeReceiver_loop:rCoObsnt##00Rcv parsing message DEBUG MIGInterpreter_parse:SECURE_RTPS_PREFIX from 0XDFCD91E1,0X6868BAE7 DEBUG MIGInterpreter_parse:INFO_TS from 0XDFCD91E1,0X6868BAE7 DEBUG MIGInterpreter_parse:DATA from 0XDFCD91E1,0X6868BAE7 At this point, the SensorPublisher_1 application is generating log messages at the DEBUG level as shown in the terminal window, but the log messages are not being forwarded to *Observability Collector Service* because the Forwarding Verbosity is still at WARNING. To set the logging Forwarding verbosity to DEBUG, repeat steps 4 - 5 above using the Log Fowarding Verbosity pull down. After setting both the Collection and Forwarding verbosity to DEBUG, you should see DEBUG messages for the SensorPublisher_1 application in the Log Dashboard. To view the Log Dashboard, click the Home icon to go back to the Alert Home dashboard. .. figure:: static/grafana_home_icon.png :figwidth: 100 % :alt: Grafana home icon :name: FigureGrafanaHomeIcon :align: center .. note:: If you are using the dashboards as an admin user, you will be prompted to save your changes. Select the **Discard** button; the changes to the dashboard do not need to be saved, since they are set in the application. This prompt does not appear when running with lesser permissions. .. figure:: static/grafana_unsaved_changes_dialog.png :figwidth: 100 % :alt: Grafana unsaved changes dialog :name: FigureGrafanaUnsavedChangesDialog :align: center Now that both Collection and Forwarding verbosity are set to DEBUG, you can examine the DEBUG log messages. Note that there are no additional logs indicated in the Warnings and Errors Logs indicators. .. figure:: static/dashboard_warnings_errors_indicators.png :figwidth: 100 % :alt: Grafana dashboard warnings errors indicators :name: FigureDashboardWarningsErrorsIndicators :align: center From the Alert Home dashboard, select the **Logs** indicator to open the Log Dashboard. .. figure:: static/dashboard_log_panel_selection.png :figwidth: 100 % :alt: Grafana dashboard log panel selection :name: FigureDashboardLogPanelSelection :align: center Verify that DEBUG messages are shown in the Log Dashboard. .. figure:: static/grafana_debug_logs.png :figwidth: 100 % :alt: Grafana debug logs :name: FigureGrafanaDebugLogs :align: center You can manipulate the Log Control settings to verify application and dashboard behavior as shown in :numref:`TableLogVerbosityDebugBehavior`. .. list-table:: Collection and Forwarding Log Verbosity DEBUG Behavior :name: TableLogVerbosityDebugBehavior :widths: 25 25 25 25 :header-rows: 1 * - Collection Verbosity - Forwarding Verbosity - Application DEBUG Log Output - Grafana Connext DEBUG Logs * - WARNING - WARNING - NO - NO * - DEBUG - WARNING - YES - NO * - WARNING - DEBUG - NO - NO * - DEBUG - DEBUG - YES - YES Close the Applications ------------------------- When done working with the example, enter ``quit`` in each running application to shut it down.