Skip to content
Halogas Ijul (aka Hilmy Izzulhaq) edited this page Nov 20, 2026 · 1 revision

R&D

Home › Using Granary › Work together › R&D

On this page: AI in this module · Find your way around · Product lines · Create an R&D project · Work in a project · The backlog across projects · Connect GitLab · Two-way issue sync · Settings that affect this module · Permissions

The R&D module tracks product development: a shared catalog of product lines, R&D projects with a backlog, sprints, releases and a roadmap, and an optional link to a self-hosted GitLab server. With the link on, the backlog and the GitLab issues of a repository stay in step both ways, and each project shows its build (pipeline) status. Engineers and product owners use it; whoever runs the GitLab server sets up the link once.

R&D projects list


AI in this module

  • Tool ideas from Nary: when you ask Nary for something none of its tools can do, it says so and files the idea in the R&D backlog of the project named in nary.skills.tool_ideas_project, labelled nary-idea. A repeated idea adds its example to the open item. Nary never writes or changes code. The setting is empty by default, so nothing is filed until you name a project.
  • Ask AI about an attachment: on an R&D project's PDF or image attachment, click the sparkle button (Ask AI about this document).

See AI Across Granary.


Find your way around

Open R&D in the sidebar:

Page What it shows
Projects Every R&D project with its product, status, latest version and items done
Backlog All backlog items across all R&D projects
Product Lines The shared catalog of product lines

The GitLab link is set up under Settings > GitLab Integration.


Product lines

Product lines are a shared catalog used across R&D, manufacturing and products. For Kumo Trading Co. these could be "Office furniture", "Printers and copiers" and "Internal tools".

  1. Open R&D > Product Lines and click New product line.
  2. Enter a Code, a Name, a Category, a Color and a Description, and keep Active on.
  3. Click Create.

Anyone signed in can read the catalog. Only the Owner or the Executive role can add, change or delete product lines. Inactive lines are hidden from pickers.


Create an R&D project

  1. Open R&D > Projects and click New project. The New R&D project dialog opens.
  2. Enter a Project name (for example "Order portal v2"), a Code and, if it belongs to one, a Product (a product line).
  3. Add a Description of the scope and goals.
  4. Click Create project. The code and the name are required.

The code is stored in capitals and must be unique: a code that is already in use is refused. The list shows the tiles Total projects, Active, Total items and Products. Use the tabs (All projects, Active, Paused, Archived) and Search projects to narrow it, then click Open project.


Work in a project

The project page shows the status and code, the latest released version, and a button per linked GitLab repository with its last pipeline status (Passed, Failed, Running, Pending). If a sprint is running, a banner shows its name, goal, end date and a Complete sprint button.

R&D project with the backlog board

The tabs are Backlog board, Sprints, Releases, Roadmap and Files.

Backlog board

The backlog board has the columns Backlog, Refined, In Sprint, In Progress, Review and Done. It works like the Projects board: drag cards between columns, filter by assignee, priority and epic, manage Epics and Columns, and click a card to open its drawer with Comments (use @ to mention someone), Activity and Files.

To add an item with details:

  1. Click Add item.
  2. Enter a Title, then pick a Type (Feature, Bug, Tech debt, Research, Spike, Docs, Security or Internal tool) and a Priority (Low, Medium, High, Critical).
  3. Optional: enter Points (story points), pick a Sprint (or leave it unassigned in the backlog), and add Labels, comma separated.
  4. Click Add item.

Completed or released items cannot be deleted.

Sprints

  1. On the Sprints tab, click New sprint.
  2. Enter a Sprint name, a Sprint goal, the Start date and End date, and a Velocity target (points).
  3. Click Create sprint. The sprint starts in planning.
  4. Click Start on the sprint when work begins. Only a sprint in planning can be started.
  5. When the sprint ends, click Complete (or Complete sprint on the banner). Enter the Velocity achieved (points) and the Retrospective notes, then click Complete sprint.

When you complete a sprint, every item in it that is not Done goes back to the backlog as Refined, ready to plan again. Only an active sprint can be completed.

Releases

  1. On the Releases tab, click New release.
  2. Enter the Version (required, for example 2.1.0), a Release date, a Release name and Release notes (Markdown).
  3. Click Create release.
  4. Click Publish when the version ships. The release becomes released, gets today's date if it had none, and becomes the project's latest version.

A release that carries a GitLab tag shows a button that opens the tag.

Roadmap and files

The Roadmap tab shows the project's roadmap notes. The Files tab holds attachments for the project.


The backlog across projects

R&D > Backlog lists all items from all projects with Item, Type, Priority, Points and Status. Filter by project and priority, or search by title. Items linked to GitLab have a link that opens the GitLab issue.


Connect GitLab

You need rnd:manage for every step. Open Settings > GitLab Integration. The How this works box on the page repeats these steps.

1. Add the server

  1. In GitLab, create a personal access token with the api scope (your avatar > Edit profile > Access tokens).
  2. In Granary, click Add server. Enter a Display name, the GitLab URL (for example https://gitlab.example.com) and the Personal access token, then click Add server.
  3. Open the server (Edit), click Generate next to Webhook secret, save, and copy the secret.

Until a webhook secret is set, Granary refuses every GitLab webhook and the page warns you.

2. Link repositories

  1. Click Add repository.
  2. Pick the project from the server's list (or click Enter the project by hand instead and fill in the GitLab project ID, Full path, Web URL and Default branch).
  3. Choose its R&D project and save.

Click Sync on a repository to pull its recent pipelines. After that, GitLab reports new pipelines by itself.

3. Add the webhook in GitLab

In each GitLab project, open Settings > Webhooks > Add new webhook:

  • URL: the webhook URL shown in the server's details (use Copy webhook URL).
  • Secret token: the webhook secret from step 1 (use Copy secret). One secret serves every project on that server.
  • Trigger: tick Push events, Tag push events, Comments, Merge request events, Issues events and Pipeline events.
  • Keep Enable SSL verification on, then click Add webhook.

Check it with Test > Pipeline events: GitLab should show HTTP 200. A 401 means the secrets differ, or no secret is set in Granary.

GitLab Integration settings with a server and linked repositories


Two-way issue sync

Turn on Sync backlog with GitLab issues on a linked repository. Granary asks Turn on the issue sync?; click Turn on and import.

  • First sync: every GitLab issue becomes a backlog item, and every backlog item gets an issue (Done items as closed issues). You get a notice with the numbers.
  • What follows: title, description with acceptance criteria, status, priority, type, labels, assignee and comments. Story points, sprint, epic and board order stay in Granary only.
  • Status: Done closes the issue. Other statuses show as GitLab labels such as status::in_progress, which a GitLab issue board can use as columns.
  • Both sides changed: the newer edit wins. The other value is kept in the item's history.
  • Deleting: an item deleted in Granary closes its issue with the label erp::deleted. An issue deleted in GitLab leaves the item in Granary, unlinked.
  • People: users are matched by email where GitLab shows it. Set the rest by hand in the People section. Comments written in Granary are posted by the access token's account, as "Name (ERP): ...".
  • Timing: with the Issues events and Comments triggers on, changes arrive at once. Granary also pulls recently updated issues on a schedule, so a missed webhook is caught.

If GitLab refuses a change, Granary retries it later and shows Last error on the repository, and a count of failing items. Click Sync issues now to run the sync at once. The sync is refused when it is off for the repository, the repository has no R&D project, or its server has no token.


Settings that affect this module

Under Settings > General, area Integrations & access, block GitLab:

Setting Key What it does
Issue sync: items per minute rnd.gitlab.sync_batch Backlog items and comments sent to each repository per minute (default 50); the rest go the next minute
Issue sync: pull every rnd.gitlab.pull_minutes How often recently updated issues are read from GitLab (default 15 minutes)
Issue sync: longest retry wait rnd.gitlab.retry_max_minutes A refused item is retried after 1, 2, 4 ... minutes, up to this (default 60)

The GitLab servers, tokens, webhook secrets and repositories are set on Settings > GitLab Integration. The access token and webhook secret are stored encrypted.


Permissions

Permission Allows
rnd:view See R&D projects, sprints, backlog, releases, linked repositories, pipelines and GitLab events; comment on items
rnd:manage Everything in rnd:view, plus create and change projects, sprints, items, releases, board columns and epics, and set up GitLab

Product lines are edited only by the Owner or the Executive role. The GitLab webhook itself needs no user: it is checked against the webhook secret.


Related pages


← Previous: Projects | Next: Knowledge and Documents →

Clone this wiki locally