Network and Storage Defaults

Most users can change their Network & Storage defaults to reduce unnecessary credit burn with minimal material impact to their pipelines.

Here’s our recommended defaults for most users:

*Workspaces may need to be increased due to some nuanced but nonetheless significant considerations. See below!

The impact of reducing Artifacts from 30 days to 7 days, simply means that artifacts will only be available 7 days from creation instead of 30.

Common use cases for artifacts are additional debug logs, test reports, or generated assets tend to have a limited shelf-life of usefulness. Clearly, if folks need to access such artifacts beyond 7 days from creation, they’ll need to increase this.

This is the most nuanced configuration option. The only times workspaces are re-used, is when either of the 2 conditions occur:

• A workflow is “on hold” for longer than the workspace retention period. Downstream jobs may then fail because the workspace no longer exists

We should explicitly check whether this is a use-case of the customer before recommending a 1-day retention period.

• A user retries a workflow “from failed” beyond the workspace retention period. This workflow may then fail, as the workspace no longer exists

The main impact of reducing cache retention from 15 days to 7 days, is that the customer may see slower builds more frequently as a result of hitting a cold cache.

We think that most folks consider a weekly cold cache as being acceptable. However given the cadence of builds through a typical project pipeline, many teams end up creating new cache entries more than once per week (e.g updating dependencies), and they are unlikely to see any material impact here at all.