Artifacts

Artifacts is the tool behind Artifact History. It is one list of everything in the tenant, all artifact types and not only integration flows, and a record of how each artifact changed. It replaces the Artifact Cache of version 1.

Turn it on

“Refresh from tenant” syncs on demand and shows its progress; the first sync of a large tenant takes a few minutes. For a history that does not depend on someone pressing a button, turn on the schedule under Artifacts → Settings. It is off by default; the default interval is one hour and the shortest is 15 minutes.

Flow Image, Flow Diff, Credentials Where-used, Flow Dependency, Flow Schedules and Deployment Status all read what this sync collects.

The inventory

• Search by name, ID, description or author, and narrow by package, type or adapter type. The filtered view is a link you can share.
• See which artifacts are drafts, and what version a draft was last saved as. Drafts block package export and transport.
• See what is deployed next to the design-time version, and which deployable artifacts are not deployed.
• Search the content of every script, mapping and flow of the tenant for a field, a URL or a function name. Orders_ToS finds Orders_ToSAP.

One artifact

• Overview, the diagram of a flow, and its files, to read without downloading and unzipping, including the files inside a service interface.
• Download as a zip you can import into a package. This also works for an artifact that was deleted on the tenant.
• History: every recorded state, with who saved it where Cloud Integration reports that. Compare any two states and see the real change, not the reordering the flow editor does on every save. Pin a draft state you may need again.

Changes

“What changed on this tenant since Monday?” The Changes page lists everything that was added, changed or deleted in a time range, for a look before a release or after an incident.

What is kept

Saved versions, the first state, renames, moves between packages, pinned states and the current state of every artifact are kept forever. Only states of unsaved work are thinned over time: by default every draft is kept for 30 days, then the last of each day for 90 days, then the last of each week for a year. The numbers are settings per tenant, and the settings page shows what the history costs on disk.

The history cannot be fetched again from the tenant, so include it in your backup.

Community and enterprise

The sync runs with every license, on a schedule too. The community license opens the current state and the latest change of each artifact; older states are listed and marked as locked. The enterprise license opens every state: compare any two and download old versions. Nothing is dropped when a license ends.

Good to know

• The history is as fine as the schedule. Cloud Integration only serves the current working copy. With the hourly sync, a draft saved thirty times in an afternoon is recorded a handful of times, and a version saved and edited again between two syncs is marked as a gap.
• Only integration flows report who changed them and when. For other types the history says when Companion saw the change. An artifact changed through the API shows the OAuth client as its author.
• Content search covers the current state, not the history, and skips binaries and text files over 4 MB.
• Not covered: integration adapters, function libraries, and OData, REST and SOAP API artifacts, which the public API does not list.
• Line numbers in the comparison of a flow model refer to the sorted XML.