Migrate from JBCF Connect to JBCF Forge

JBCF Forge is the Forge edition of Calculated Fields for Jira Cloud. It replaces JBCF Connect (the legacy Connect edition) on the same Marketplace listing: you move with the Update button in Manage apps, and you do not buy the app again. This guide explains what changes when you move, how your Connect settings carry over, and what you redo by hand.


At a glance

Area JBCF Connect JBCF Forge
Marketplace listing Calculated and other custom fields (JBCF) The same listing - installed with Update in Manage apps
Platform Atlassian Connect + vendor-managed server 100% Forge native - Runs on Atlassian
Hosting Heroku-backed vendor service Atlassian-hosted Forge runtime
Storage Vendor-managed database Field values in Jira; settings in Forge KVS (Atlassian-managed, encrypted at rest)
Outbound network Required - Connect callbacks to a vendor URL None - no external: manifest scopes
Sub-processors Heroku and Atlassian Atlassian only
Monitoring Vendor logs and status page Forge app-monitoring (Forge App Logs API)
Field set 29 derived fields 31 calculated fields plus the interactive ToDo List
JQL values of the Yes/No fields 'true' / 'false' Yes / No
User fields Store the user’s name User-typed fields that hold the accountId
Recalculation on migration n/a Values are recalculated from issue history, not copied

What changes

Where the app runs

JBCF Connect is a Connect app: a vendor-managed server (running on Heroku) receives Connect callbacks from Jira and reads issue data via the Jira REST API. JBCF Forge runs entirely inside Atlassian Forge - every resolver, event handler, and scheduled job executes on Atlassian’s serverless runtime, next to your Jira instance. There is no vendor server to operate.

That is what the Runs on Atlassian badge on the Marketplace listing represents: the app’s processing and storage stay inside Atlassian’s perimeter.

Where data lives

Connect persists app data in a vendor-managed database. Forge keeps its settings in Forge KVS, an Atlassian-managed key-value store that is encrypted at rest and isolated per installation. Calculated values exist only as the value of the Jira field, so JQL and Issue Navigator searches index the up-to-date value.

How data migrates

Values are not copied from Connect - JBCF Forge recalculates them from issue history. Forge fields are new fields with new ids, and the app does not create fields itself: a Jira administrator creates them, and the app links each one to its Connect settings and fills in its values. Your Connect settings carry over through the Migration from Connect tab of the app administration:

  1. Update the app. In Manage apps, click Update on the JBCF listing. The Forge edition replaces the Connect edition; you do not buy it again, and the two editions cannot run at the same time. Before you update, download the settings file from Connect as a fallback (see If settings did not move automatically).
  2. Open the checklist. The app looks for the Connect settings by itself every hour, so nothing has to be exported by hand. The Migration from Connect tab lists Fields to create: for each Connect setting, the exact Name, Field type and Projects × issue types the Forge field needs.
  3. Create the fields. As a Jira administrator, create each field in Jira settings → Issues → Custom fields with the exact name and type from the checklist, and add the listed projects and issue types to the field’s context.
  4. Click Run again. Fields are not linked automatically. The app links each new field to its Connect setting by an unambiguous type + name pair (the name match ignores case) and queues a recalculation of the values from issue history. The page reports “Linked N new field(s); M recalculation task(s) queued.” Progress is shown on the Service Info tab (see Recalculation).

The Migration from Connect tab

Part of the tab What it shows
Settings from Connect found N, not found, not readable (HTTP …) or being written, check again in an hour. If the settings are not found, upload the file from Connect (below).
Upload file from Connect The manual fallback, see If settings did not move automatically.
Status Done, Fields to create or Pending, shown only when the Connect settings or the uploaded file are found: it describes them, not an earlier pass.
Last migration When an earlier pass was saved, a separate line with the date and time of the last pass (automatic, Run again or file upload): “Last migration on …: N field(s) linked, M to create, K recalculation task(s) queued.” Running the same settings again updates the date. A pass saved before this version has no run date, so the line shows the date the settings were exported from Connect instead: “Last migration (settings exported from Connect on …): …”. It stays on the tab when the settings are no longer found, for example after a move by file.
Fields to create The checklist of fields to create. A setting that was limited by JQL in Connect is marked “Selected by JQL in Connect: use the projects and issue types it covers”.
Run again Links the fields you created. While fields are still to create, a line under the Run again and Upload file from Connect buttons says so: “After you create the fields from the checklist, select Run again: they are not linked automatically.”
Linked fields Connect settings already linked to a Forge field.
Needs your choice A link is made only when the pair “type + name” is unique on both sides. If several Connect settings or several Forge fields share a name and type, rename the extra ones, then Run again.
Not migrated Types that the Forge edition does not have yet: “the type … is not available in the Forge edition yet. Contact support if you need it.”

The tab also warns you about differences between the editions:

  • The JQL filter of a Connect field is not carried over: Forge fields have no JQL filter, the value is calculated on every issue of the field’s contexts.
  • User fields stored the user’s name as text in Connect; in Forge they are user-typed fields that hold the accountId. Rewrite filters that compare or search them.
  • The field’s contexts do not cover all the projects and issue types it had in Connect - the warning lists what to add.

If settings did not move automatically

When the tab says not found or not readable - the Connect edition never wrote its settings, Connect is already removed, or support asked you to - move the settings with a file:

  1. In the Connect edition, open the app’s admin page, tab Calculated custom fields(JBCF), and select Download for Forge (or get the file from support). Only a Jira administrator can download it; the file is jbcf-forge-migration.json. Download it before you click Update, while the Connect edition is still installed.
  2. On Migration from Connect in the Forge edition, select Upload file from Connect and choose that file.

The file goes through the same steps as the automatic move: the line reads “Settings from the uploaded file: found N”, the same checklist, linked fields and warnings appear, and the page reports “Linked N new field(s); M recalculation task(s) queued.” The file is not kept by the app.

Fields you create after the upload are not linked by the hourly check, because the app does not keep the file: “Create the missing fields, then upload the file again.” Uploading the same file again changes nothing that is already linked.

A file that cannot be used is refused with the reason:

  • “This file is for another app.” - the file was downloaded from a different Connect app;
  • “The file is damaged: its checksum does not match its contents. Download it from Connect again.”;
  • “The file is not valid JSON. Upload the file downloaded from Connect without changes.”;
  • “This is not a migration file from Connect.”;
  • “The file holds no field settings. Download it from Connect again.”;
  • “The file is too large: the limit is 5 MB.”

Filters after the move

Field IDs change with the move, so saved filters that use cf[id] of an old Connect field stop matching. On the Migration from Connect tab, Show filter audit opens the read-only Filters after the move from Connect report. It:

  • lists the Connect fields Jira still shows;
  • pairs each with the new Forge field by name (Paired by name, Several fields with this name, No Forge field with this name);
  • lists the filters that refer to an old field by cf[id] or by its name, with the current JQL and a Suggested JQL (Copy JQL).

A part with nothing to show says so: “Jira shows no fields of the Connect edition.”, “No Connect fields to pair with new fields.”, “No saved filter uses an old Connect field.” Hide filter audit folds the report; Refresh reads it again.

In the suggested JQL, for user fields the user’s name is replaced with the accountId, and the in, not in, and != operators are handled; where the suggestion needs a user, the report says to replace the placeholder with the right account id. A field name that matches a system field (Assignee, Reporter, and so on) is not replaced - the report shows a warning instead.

For the four Is author of first / last comment assignee / reporter fields the suggestion also changes the value: Connect stored true / false, the Forge edition stores Yes / No, so = true becomes = "Yes" and = false becomes = "No" (~ becomes =). One difference stays: Connect left these fields empty on an unassigned issue (or one without a reporter), the Forge edition shows No - a filter with = false now finds such issues too, and one with is EMPTY no longer does.

The report does not change filters. Only the owner of a filter can edit it, so copy the suggested JQL and send it to the owner, who applies it to their own filters.

How the field libraries compare

JBCF Connect ships 29 derived fields. The Forge edition covers the same derived-field territory - comments, changelog, links, project context, hierarchical sort - adds SLA breach time for Jira Service Management, and adds the interactive ToDo List field, for 31 calculated fields plus ToDo List.

Formula fields, picker fields (status, role, group, version), and utility fields (regex, time slider, attachment select) exist in the Data Center edition of JBCF. They are not part of either Cloud edition; see the feature difference table for the Data Center to Cloud picture.

What you will have to redo by hand

  • Creating the fields. The checklist on the Migration from Connect tab tells you which fields to create; a Jira administrator creates them with the exact name and type, sets the projects and issue types in the field context, and attaches them to screens.
  • Fields in Needs your choice and Not migrated. Resolve the duplicated names; for types the Forge edition does not have, decide what replaces them.
  • Saved filters. Each filter owner applies the suggested JQL from the Filters after the move from Connect report. Filters on a name that matches a system field need a manual edit.
  • JQL filters on the Yes/No fields. The Connect edition stores those values as 'true' / 'false'; the Forge edition stores Yes / No. Saved filters and boards that compare against 'true' or 'false' have to be updated.
  • Field IDs. The Forge fields get new IDs, so any automation rule or integration that refers to a field by ID needs the new ID.

What stays the same

  • Behavior of derived fields. The derived fields produce the same logical value as their Connect counterparts: cf-number-of-comments counts comments, cf-previous-status returns the prior status, cf-sort-by-epic-rank produces the same composite sort key.
  • Searchability. Each calculated value is written back to Jira as a standard custom field, so JQL filters, Issue Navigator columns, list views, and Export to CSV all keep working - after you point them at the new fields.
  • Read-only model. Calculated fields remain read-only - their value is always derived from issue state. ToDo List is the one field users edit.

Why move

The Forge edition is for you if you:

  • Want the Runs on Atlassian posture (data and processing inside Atlassian).
  • Need to satisfy compliance reviews that flag external sub-processors or vendor-hosted infrastructure.
  • Want the derived fields the Connect edition offers, plus the interactive ToDo List field.
  • Want lower operational risk - no companion server, no Heroku dependency, no vendor outage to wait through.

The Connect edition is in maintenance: it is kept working and supported, but new capability goes into the Forge edition. Atlassian ends support for Connect apps on 31 January 2027, so the move is a matter of when, not whether.

The two editions cannot run at the same time: the Forge edition is an update of the same app, not a second app. Plan the move as one step - update, create the fields from the checklist, click Run again, then fix the saved filters from the report. On the day the Forge edition is published, the Connect administration shows a Forge edition is available banner that links to the Connect to Forge page.


Adjacent JiBrok products

JBCF Forge is one app in a portfolio of Forge-native JiBrok products. Each product covers a different territory; there is no feature overlap:

  • JBCF Forge - calculated/derived custom fields (this app).
  • Time in Status (Forge) - all time-tracking metrics: status time, stopwatches, timers, work calendars, activity reports.
  • JiBrok Studio - formula scripting language and scheduled jobs.
  • Display Linked Issues - extended visualization of issue links.

Formula evaluation, time-in-status calculations, and rich link panels are not part of JBCF on Cloud in either edition - they live in the products listed above.


Next steps

If you need help planning the move or you have a Forge-edition feature request, contact JiBrok support.