Installation and configuration
You can install Python locally and the required packages or you can use Docker.
For Mac M1 users, if Pandas is not already installed, we advise you to use Docker. The current release (Spring 2022) of Pandas has some installation problems on Mac M1.
1️⃣ Clone the GitHub repository
2️⃣ Edit the configuration of the data observability agent
The agent collects information inside the application to send it to Kensu. The agent augments the program library and gathers metadata, defines the lineage, and computes the metrics before sending this to Kensu through the API. To configure the agent, a configuration file is needed.
Edit the /conf.ini file, with the following:
- kensu_ingestion_url: This is the URL of the API of the Kensu platform, which is already pre-configured for you if you are using the Community edition
- kensu_ingestion_token: This token will be used in the HTTPS communication with the kensu_ingestion_url in order to add an identity to the traces, logs and metrics sent with your user for easier governance and management
- kensu_api_token: This is the Kensu API (Personal Access Token) which will allow you to Create a monitoring rule programmatically
If you haven't got your kensu_ingestion_url and kensu_api_token yet, please follow the instructions in Getting Kensu Credentials.
3️⃣ For Docker Only
If Docker is not installed, please follow these steps.
(The minimum requirements for Docker are 4 CPUs, 4 GB of Memory, 1GB of Swap, 40GB of Disk Image.)
Run this command to build the image:
4️⃣ For Python/PySpark Installed Locally
- Create a python virtual environment: python3 -m venv kensu_env
- Activate the environment: source kensu_env/bin/activate
- Install the required Python packages: pip install -r requirements.txt
Go to the next steps
