Merge pull request #8760 from MicrosoftDocs/main

Publish to live, Tuesday 4 AM PST, 6/3

Author: ttorble

docs/data-engineering/fabric-notebook-selection-guide.md

@@ -35,7 +35,7 @@ Use this high-level reference to quickly determine the most suitable notebook ty
 
 Use this structured comparison table to understand the architectural and operational trade-offs between notebook types. Best suited for engineering evaluations or implementation planning. 
 
-**Execution & Performance**
+### **Execution & Performance**
 
 | Scenario | Python Notebooks (2-core VM) | PySpark Notebooks (Spark Compute) |
 | --- | --- | --- |
@@ -46,27 +46,27 @@ Use this structured comparison table to understand the architectural and operati
 | High-Concurrency Execution | Manual FIFO-style parallelism per notebook | System-managed concurrency with support for parallel execution. |
 | Resource Customization & Scaling | Fixed compute (2-core VM); does not auto scale. Users can manually scale out using %%config within the notebook. | Flexible resource allocation; supports autoscaling and custom Spark configurations. |
 
-**Workflow & Orchestration**
+### **Workflow & Orchestration**
 
 | Scenario | Python Notebooks (2-core VM) | PySpark Notebooks (Spark Compute) |
 | --- | --- | --- |
 | API Orchestration | Effective for lightweight orchestration and control flows, especially REST/gRPC-based integrations | Less optimal for basic orchestration tasks due to longer start-up and distributed overhead. |
 | Complex ETL DAGs | Supports FIFO scheduling only | Supports both FAIR and FIFO scheduling with DAG customization. |
 
-**Platform & Library Support**
+### **Platform & Library Support**
 
 | Scenario | Python Notebooks (2-core VM) | PySpark Notebooks (Spark Compute) |
 | --- | --- | --- |
 |  Library Access | Strong Python library support across multiple runtimes; however, limited access to Spark-native libraries may require manual integration. | Fully supports MLlib, Spark SQL, PySpark, and Spark Streaming. |
 | Delta Lake Handling | The Python Notebook runtime comes with pre-installed [deltas](https://delta-io.github.io/delta-rs/) and [duckdb](https://duckdb.org/) libraries, enabling both reading and writing of Delta Lake data. However, some Delta Lake features may still be unsupported. | Fully supported with native compatibility. |
 
-**Production & Enterprise Readiness**
+### **Production & Enterprise Readiness**
 
 | Scenario | Python Notebooks (2-core VM) | PySpark Notebooks (Spark Compute) |
 | --- | --- | --- |
 | Production Management | Limited production features; Does not support environment vars. | Support for production workflows with environment variables, library management through environment items, and item-based deployment. |
 
-**Cost Considerations**
+### **Cost Considerations**
 
 | Scenario | Python Notebooks (2-core VM) | PySpark Notebooks (Spark Compute) |
 | --- | --- | --- |
@@ -80,15 +80,15 @@ Use **Python Notebooks** for fast iteration, cost-effective analysis, and intera
 
 Use **PySpark Notebooks** for distributed computing, production-grade ETL workflows, or scenarios where high concurrency and Spark-native APIs are essential. 
 
-**Choose Python Notebooks When:**
+### **Choose Python Notebooks When:**
 
 - You need fast start-up (typically within seconds) on a lightweight 2-core container. 
 - If minimizing compute cost is a priority - for interactive analysis or scheduled micro-jobs. 
 - You want immediate access to pip-installable libraries and pre-installed DuckDB and Polars. 
 - You need to test across different Python runtime versions. 
 - Your data comfortably fits in the memory of a single node. 
 
-**Choose PySpark Notebooks When:**
+### **Choose PySpark Notebooks When:**
 
 - Your workloads exceed the memory or compute limits of a single node. 
 - You require high-concurrency pools to run parallel jobs across Notebooks. 
@@ -109,8 +109,6 @@ This section provides a quick reference for the fundamental technical and archit
 | Cost Profile | Lower cost; ideal for short tasks and prototyping | Higher cost; suited for scalable, long-running workloads |
 | Python/Spark Versioning | Multiple Python versions available | Tied to specific Spark runtime version |
 | Custom Libraries | pip install + resource folders | pip install + resource folders + environments item |
-| ~~Environment Variables~~ | ~~Not supported~~ | ~~Supported~~ |
-| ~~Lakehouse Integration~~ | ~~Requires Lakehouse to be explicitly attached~~ | ~~Can read/write without attachment~~ |
 | Fabric Spark Capabilities | Limited access to Spark engine features | Full access: NEE, Autotune, VORDER, Vegas Cache |
 | Delta Lake Compatibility | Partially compatible; potential performance issues | Fully supported and optimized |
 

docs/data-engineering/runtime.md

@@ -32,7 +32,7 @@ Below, you find a comprehensive comparison of key components, including Apache S
 
 |                       | **[Runtime 1.1](./runtime-1-1.md)** | **[Runtime 1.2](./runtime-1-2.md)** | **[Runtime 1.3](./runtime-1-3.md)** |
 |-----------------------|-------------------------------------|-------------------------------------|------------------------------------|
-| **Release Stage**     | EOSA                                | GA                                  | GA                                 |
+| **Release Stage**     | EOS                                    | EOSA| GA                                 |
 | **Apache Spark**      | 3.3.1                               | 3.4.1                               | 3.5.0                              |
 | **Operating System**  | Ubuntu 18.04                        | Mariner 2.0                         | Mariner 2.0                        |
 | **Java**              | 8                                   | 11                                  | 11                                 |