Databricks and Logging in Notebooks
0 likes542 views
The document discusses etiquette for attending Databricks sessions, emphasizing punctuality, feedback submission, and minimizing disturbances. It also provides an overview of Azure Databricks, detailing its features like managed Spark clusters, notebooks, and logging systems. Additionally, it highlights the importance of logs for debugging and monitoring jobs in Databricks Scala notebooks.
![Spark Jobs work with large amount of data and the tasks involve time consuming
computations.
They run at remote locations too. So, it becomes difficult to track the execution step without
logs.
Logs help to track at what point is the execution at, they help the developer to debug at what
points is the job consuming maximum of it’s time.
Logs often contain vast amounts of metadata, including date stamps, logger name [that can be
set to be the name of the Logging Class], source information such as cluster name. This data
helps in the debugging process.
Using Real-time Logs and messages logged in them, certain Alerting techniques can be used
to create notifications if a log containing a particular message appears.
Importance of Logs](https://image.slidesharecdn.com/knolx-databricksandlogging-200922072406/85/Databricks-and-Logging-in-Notebooks-13-320.jpg)
Databricks and Logging in Notebooks
- 1. Presented By: Swantika Gupta Software Consultant Databricks and Logging in Notebooks
- 2. Lack of etiquette and manners is a huge turn off. KnolX Etiquettes Punctuality Respect Knolx session timings, you are requested not to join sessions after a 5 minutes threshold post the session start time. Feedback Make sure to submit a constructive feedback for all sessions as it is very helpful for the presenter. Silent Mode Keep your mobile devices in silent mode, feel free to move out of session in case you need to attend an urgent call. Avoid Disturbance Avoid unwanted chit chat during the session.
- 3. Agenda What is Databricks? Reasons to use Azure Databricks Azure Databricks Core Artifacts Logging in Scala Notebooks Workspace Clusters Notebooks Libraries Jobs Data
- 4. What is Databricks ? Industry - leading, Zero-management cloud platform built around Spark. Delivers - fully managed Spark clusters - an interactive workspace for exploration and visualization - a production pipeline scheduler - a platform for powering your favorite Spark-based applications So instead of tackling headaches like setting up infrastructures, creating data backup, scaling your nodes according to load, you can finally focus on finding answers that make an immediate impact on your business. It is a product offered by a third - party, but it is offered as a first - class service tightly integrated with AWS and Azure
- 5. Reasons to use Azure Databricks Familiar Languages and Environment Higher Productivity and Collaboration Easy integration with Microsoft Stack Extensive List of Data Sources Suitable for Small Jobs too Extensive Documentation and Support Available
- 6. Azure Databricks Core Artifacts
- 7. Azure Databricks Artifact - Workspace An environment inside Databricks service with access to all your databricks resources Organizes various objects, like Notebooks, Jars into Folders. Provides easy one-click access to computational resources, like clusters and data stored
- 8. Azure Databricks Artifact - Clusters Core Component of Databricks A set of computational resources and configurations Runs our Data Engineering, Data Science and Data Analytics workloads Types of Clusters: - Interactive Clusters - Automated Clusters
- 9. Azure Databricks Artifact - Data ● Create Tables directly from imported data. The table schema is stored in the internal databricks metastore ● Use Apache Spark commands to read data from supported data source ● Import data into DBFS and use the DBFS CLI, DBFS API, DBFS utilities, Spark APIs, and local file APIs to access the data.
- 10. Azure Databricks Artifact - Notebooks Web-Based interface to a Document containing - Runnable Code - Visualizations - Narrations Support for multiple languages in the same notebook Real-time collaboration on the same notebook Revision History of the notebook
- 11. Azure Databricks Artifact - Jobs As an alternative to running notebooks interactively, you can set for a notebook or Jar to either run immediately or on a scheduled basis. Three types of task can be run as jobs: - Notebooks - JARs - Spark Submit Notebook Job and JAR jobs can be configured by passing parameters too Spark-Submit can also be configured
- 12. Azure Databricks Artifact - Libraries A Library can be installed on a cluster to make some third-party or custom code available to the running notebook or JAR Libraries can be installed in 3 modes: - Workspace Libraries - Cluster Libraries - Notebook scoped Libraries
- 13. Spark Jobs work with large amount of data and the tasks involve time consuming computations. They run at remote locations too. So, it becomes difficult to track the execution step without logs. Logs help to track at what point is the execution at, they help the developer to debug at what points is the job consuming maximum of it’s time. Logs often contain vast amounts of metadata, including date stamps, logger name [that can be set to be the name of the Logging Class], source information such as cluster name. This data helps in the debugging process. Using Real-time Logs and messages logged in them, certain Alerting techniques can be used to create notifications if a log containing a particular message appears. Importance of Logs
- 14. Logging in Databricks Scala Notebooks Databricks’ Log Delivery System Delivers Spark Driver, Executor and Event Logs to a location specified during configuring the cluster Delivered every 5 minutes Incase a cluster terminates, Databricks make sure to deliver all logs up till the termination to the delivery location.
- 15. Logging in Databricks Scala Notebooks Default Log4j Properties in Databricks Two log4j.properties files: For Driver: For Executors:
- 16. Logging in Databricks Scala Notebooks Overwriting the Default Log4j Properties with init scripts Drawback - Every time configuration is to be changed, cluster restart is required
- 17. Logging in Databricks Scala Notebooks Using External Log4j configuration for your job - Create a Log4j properties file for your custom loggers and appenders - Upload the file to your required DBFS location - Use Log4j’s PropertyConfigurator object to configure using your custom log4j properties
- 19. Thank You ! Get in touch with us: Lorem Studio, Lord Building D4456, LA, USA