How to create a new sample app#
This guide provides step-by-step instructions on how you can create a new sample app by taking
Wind Turbine Anomaly Detection OR Weld Anomaly Detection sample app as a reference
by configuring the generic Time Series AI Stack. Create a copy of the
edge-ai-suites/manufacturing-ai-suite/industrial-edge-insights-time-series/apps/wind-turbine-anomaly-detection
folder and replace the folder name with the sample app name to be created.
Steps to follow#
Configuration#
1. Data Simulators/Destinations:#
Options available:
Update the OPC-UA/MQTT simulator containers (
edge-ai-suites/manufacturing-ai-suite/industrial-edge-insights-time-series/simulator) as needed to ingest the needed dataset to analyze via a CSV fileIf you have existing
OPC-UA/MQTTsimulators, you can use that too. Just adjust the compose and helm template files similar to the above existing OPC-UA/MQTT simulatorsDirectly configure
Telegrafto connect to the device acting asOPC-UAserver ORMQTT publisher/broker. Refer to the Telegraf plugin documentation for more details
2. Generic Time Series AI stack#
Data Ingestion - Telegraf
Update the
[[inputs.mqtt_consumer]]and[[inputs.opcua]]sections inedge-ai-suites/manufacturing-ai-suite/industrial-edge-insights-time-series/apps/<new-app-name>/telegraf-config/Telegraf.conf, based on the MQTT/OPC-UA configuration, the input would be read and the topics in MQTT input plugin and name_override in OPC-UA input plugin is used for writing this as measurement in InfluxDB.In the above same conf file, the two
[[outputs.influxdb]]plugins publish the data to InfluxDB and Time Series Analytics microserviceData Processing - Time Series Analytics Microservice
The
tick_scripts,modelsandudfsfolders atedge-ai-suites/manufacturing-ai-suite/industrial-edge-insights-time-series/apps/<new-app-name>/time-series-analytics-configconstitutes the User Defined Function (UDF) deployment package. Theconfig.jsonat the same location is the Time Series microservice configuration.More details at how-to-configure-custom-udf
Data Visualization - Grafana
Prepare the custom grafana dashboard (
edge-ai-suites/manufacturing-ai-suite/industrial-edge-insights-time-series/apps/<new-app-name>/grafana-dashboard.yml) which gets volume mounted in the grafana microservice atedge-ai-suites/manufacturing-ai-suite/industrial-edge-insights-time-series/docker-compose.yml.
Deployment#
Note: Adjust the
edge-ai-suites/manufacturing-ai-suite/industrial-edge-insights-time-series/Makefileto add the support for the new sample app name for both docker compose and helm deployments
For docker compose deployment, follow the Get Started guide/
For helm deployment, follow the How to deploy with helm