You can manage the global settings of your IBM Watson Pipelines such as a default error policy and default rules for node caching.
Global settings apply to all nodes in the pipeline unless they are overwritten by a node setting. To update global settings, click the Manage default settings icon 
on the toolbar. You can configure:
Setting global error policy
You can define the behavior of Pipelines when an error occurs.
- Fail pipeline on error stops the flow and initiates a error-handling flow.
- Continue pipeline on error tries to continue executing the pipeline.
Error handling
You can configure the behavior of Pipelines for error handling.
- Create custom-error handling response: Customize an error handling response. Add an error handling node to the canvas so you can configure a custom error response. The response will apply to all configured nodes to fail
when an error occurs.
- Show icon on nodes linked to error handling pipeline: An icon flags a node with an error to help debug the flow.
To learn more about error handling, see Managing pipeline errors
Setting node caches
Manual caching for nodes lets you set the default for how the pipeline caches and stores information. You can override these settings for individual nodes.
Default cache utilization frequency
You can change the following cache settings:
Caching method
Choose whether to enable automatic caching for all nodes or choose to manually set cache conditions for specific nodes.
- Enable automatic caching for all nodes (recommended)
All nodes that support caching enable it by default. Setting Creation Mode or Copy Mode in your node's settings toOverwriteautomatically disables cache, if the node supports these setting parameters. - Enable caching for specific nodes in node properties panel.
In individual nodes, you can select Create data cache at this node in Output to allow caching for individual nodes. A save icon appears on nodes that uses this feature.
Cache usage
Choose the conditions for using cached data.
- Do not use cache
- Always use cache
- Use cache when all selected conditions are met
- Retrying from a previous failed run
- Input values for the current pipeline are unchanged from previous run
- Pipeline version is unchanged from previous run
To view and download your cache data, see Run tracker in your flow. You can download the results by opening a preview of the node's cache and clicking the download icon.
Resetting the cache
If your cache was enabled, you can choose to reset your cache when you run a Pipelines job. When you click Run again, you can select Clear pipeline cache in Define run settings. By choosing this option, you are only overriding the default cache settings to reset cache for the current run and not previous runs.
Managing your Pipelines settings
Configure other global settings for your Pipelines asset.
Autosave
Choose to automatically save your current Pipelines canvas at a selected frequency. Only changes that impact core pipeline flow will be saved.
Parent topic: IBM Watson Pipelines