> ## Documentation Index > Fetch the complete documentation index at: https://docs.arc.cdata.com/llms.txt > Use this file to discover all available pages before exploring further. # Performance Settings > How to configure global Automation service settings for parallel processing and worker pools in CData Arc, and how to analyze worker pool performance. This section allows you to define global settings for the [Automation](https://docs.arc.cdata.com/26.4/self-hosted/en/getting-started/architecture.md#automation) service. ![Performance settings page](https://mintcdn.com/cdata-arc/JuVWSaGlKX7avH_C/26.4/images/settings_performance.png?fit=max\&auto=format\&n=JuVWSaGlKX7avH_C\&q=85\&s=8fd458b22db9568f8df4ca7f48b3b74f) | Setting | Description | | - | - | | **Parallel Processing** | When parallel processing is enabled, Arc allocates a pool of workers to process messages for multiple connectors in parallel, as opposed to performing processing for each connector in series. Parallel processing improves processing throughput across multiple transfers and message processing across multiple stages. The following options only appear when parallel processing is enabled. | | **Worker Pool** | Establishes the global maximum number of threads that the application can assign to all connectors (combined) at the same time. When a thread finishes its assignment in a specific connector, it is recycled back into the pool. The hardware resources available on the host machine determine the upper limit for this setting. A larger worker pool consumes more system resources while Arc is processing messages. To diagnose how your worker pool is performing, see [Analyze Worker Pool](#analyze-worker-pool). | | **Max Workers per Connector** | The default number of workers that a connector uses to process transmissions. This setting can be overridden on an individual basis on the **Advanced** tab of each connector. Connectors that use session communication such as FTP, SFTP, and OFTP default to 1 worker, since multiple files might be transferred over the same connection. | | **Max Files per Connector** | The default number of files that workers assigned to a connector process before the workers are reallocated to other connectors. This can be overridden on an individual basis on the **Advanced** tab of each connector. When a large number of connectors are configured, lower values here ensure that processing is distributed evenly across all connectors. | See [Maximizing Performance](https://docs.arc.cdata.com/26.4/self-hosted/en/getting-started/architecture.md#maximizing-performance) for more information. ## Analyze Worker Pool The **Analyze Worker Pool** button collects diagnostic logs from the worker pool for a set period of time and runs an AI analysis on them. The analysis reports on the health of the worker pool, identifies performance findings, and recommends configuration changes. Arc never applies recommendations automatically. Review each recommendation and make the changes yourself. This feature requires an AI provider. If an AI provider is not configured in [AI Settings](https://docs.arc.cdata.com/26.4/self-hosted/en/getting-started/administration/settings/ai-settings.md), the button is disabled. Hover over the info icon next to the button to see why. Only members of the [Admin role](https://docs.arc.cdata.com/26.4/self-hosted/en/getting-started/administration/settings/user-roles.md#admin-role) can start, cancel, and remove analyses. Members of the Standard and Support roles can open the pane and view previously run analyses, but cannot start, cancel, or remove them. ### Starting an Analysis 1. Click **Analyze Worker Pool** to open the **Analyze Worker Pool Sessions** pane. The pane lists any previous analysis sessions. ![Analyze Worker Pool Sessions pane](https://mintcdn.com/cdata-arc/JuVWSaGlKX7avH_C/26.4/images/settings_performance_analyze_worker_pool.png?fit=max\&auto=format\&n=JuVWSaGlKX7avH_C\&q=85\&s=acc4c55d0ee2f40525bede71154e09d0) 2. Under **New Analysis**, click **Start Analysis** to open the **Configure Analysis** pane. ![Configure Analysis pane](https://mintcdn.com/cdata-arc/JuVWSaGlKX7avH_C/26.4/images/settings_performance_configure_analysis.png?fit=max\&auto=format\&n=JuVWSaGlKX7avH_C\&q=85\&s=920a0a6cf7053b34ff52492ae0f00a3e) 3. Set the following options: | Option | Description | | - | - | | **Log Collection Duration** | How long Arc collects diagnostic logs before running the analysis. Choose 1, 2, 3, 5, or 10 minutes. The default is 5 minutes. Longer windows capture more activity but delay the results. | | **Log Analysis Verbosity Levels** | Controls how much detail appears in the results. The default is **Summary**. **Summary**: Health assessment and key limitations only. **Standard**: Adds status information and identified performance findings. **Detailed**: Full analysis including raw metrics and optimization recommendations. | > [!NOTE] > > The verbosity level is set when the analysis starts and cannot be changed afterward. To see more detail, run a new analysis with a higher verbosity level. 4. Click **Start Analysis**. The analysis runs in two phases: **Collecting logs**, which shows the time remaining, followed by **Analyzing logs**. You can close the pane while the analysis runs. When it finishes, Arc displays an in-app notification in the page header. Click **Go to Results** to open the analysis, or **Go to Details** if the analysis ended in an error. To stop an in-progress analysis, click **Cancel Analysis** and confirm. ### Viewing Results Each completed analysis is saved and listed under **Previous Analysis Sessions** in the **Analyze Worker Pool Sessions** pane, along with its file count and size. Click a session to open its **Worker Pool Log Analysis** report. Click the copy icon in the report header to copy the full analysis. ![Worker Pool Log Analysis results](https://mintcdn.com/cdata-arc/JuVWSaGlKX7avH_C/26.4/images/settings_performance_worker_pool_analysis_results.png?fit=max\&auto=format\&n=JuVWSaGlKX7avH_C\&q=85\&s=899cb8ef182ed654dc0706298ba809ef) The sections that appear in the report depend on the verbosity level you selected: | Section | Description | | - | - | | **Summary** | The overall health of the worker pool over the collection window. | | **Status** | What the analyzer read to produce this report. | | **Findings** | Behavior the analyzer identified in the collected logs. | | **Limitations** | What the collected logs were not able to answer. | | **Metrics** | Values the analyzer computed from the collected logs. | | **Recommendations** | Configuration changes the analyzer suggests. Each recommendation lists the **Setting**, its **Config Level**, the current and recommended values, the **Rationale**, and the **Expected Impact**. | | **Additional Details** | Anything else the analyzer reported that does not fit the sections above. | To delete a saved analysis, hover over the analysis and click the trashcan icon. Confirm your selection and click **Remove**. Only members of the Admin role can remove analyses. This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.