Limits and known gaps
This page lists what Spacewright does not do, or does not do yet. If something here blocks you, tell us: it helps us decide what to fix first.
Not copied
- Page history. Only the current version of each page is copied.
- Comments, both page comments and inline comments.
- Whiteboards and databases. Confluence offers apps no way to copy them. The dry run and the copy report list each one under Not copied, with where it is in the source space, including those at the top level of the space. Pages inside them are still copied.
- Smart links in the content tree (embeds). Listed under Not copied in the same way. Add them again in the new space.
- Drafts, and archived pages unless you tick the option.
- Older attachment versions. Only the latest version of each file is copied.
- Personal labels. Only ordinary (global) labels are copied.
- Space settings: space colours (look and feel), space logo, sidebar shortcuts and space categories are never copied. The description, theme and space settings are copied only when Space look is ticked; otherwise the new space starts with an empty description.
- Authors. Copied pages are created by the person who ran the copy, not by the original authors.
Sizes and numbers
| Limit | What happens |
|---|---|
| Attachments over 250 MB | Not copied; listed as failed. |
| Pages whose content is over 5 MB | Created but not filled; listed as failed. |
| Copies running at once | Two per site. Others wait in the queue. |
| Variables per template | 30, with values up to 500 characters each. |
| Rows per table in the report | 500. The report says how many more there were. |
| Jobs shown in the Jobs tab | Your last 30, from the last 30 days. Jobs are deleted 30 days after they end. |
| Admin audit log | Copies from the last 30 days (up to 100 shown). Older copies appear only as monthly totals. |
| Space list in Copy a space | Up to 250 spaces at a time, in alphabetical order; Show more spaces loads more. A search reads up to 2,000 spaces at a time; Search further continues. Personal and archived spaces are not listed. |
Spacewright runs on Atlassian's Forge platform and works through Confluence's REST APIs. Atlassian's platform quotas and Confluence rate limits decide how fast a copy can go: a large space with many attachments can take hours, and a copy pauses when Confluence asks apps to slow down.
Content at the top level of a space
Folders, whiteboards, databases and smart links can sit at the top level of a space, next to the home page, instead of under a page. Confluence's API can list a space's top-level pages, but not its other top-level content, so Spacewright finds that content through Confluence search:
- Top-level folders are copied to the top level of the new space, with everything inside them.
- Top-level whiteboards, databases and smart links are listed under Not copied with the note Top level of the space. Pages inside them are copied to the top level of the new space, after the other top-level items, and the report's Notes say so.
- Search can lag behind. Something created at the top level a few moments before the copy may not be found yet. Wait a minute after big changes to the source space, and check the dry run.
- If Confluence search refuses the request, the copy still goes on. The dry run and the report then say, at the top of their notes, that top-level items may be missing.
Known gaps we plan to fix
- Find placeholders in this space looks only at the titles of top-level pages and of pages below the home page, and at the home page's text. It does not look inside folders at the top level of the space (details).
Links Spacewright does not switch yet
Spacewright switches the common link formats: page and blog post links, anchors, attachment links, smart links, page addresses, short links, and the space or page settings of macros such as Children, Page tree, Include page, Excerpt include, Content by label and Page properties report. These cases are not handled yet:
- Placeholders in pages from the legacy editor that are broken up by more than formatting, for example by a line break, or whose name itself is partly formatted (listed in the report, not filled). Placeholders that are only wrapped in bold, italic, underline, strikethrough, colour, subscript, superscript or code formatting are filled.
- Macros from other apps whose settings hold a page title, or a space key in a setting Spacewright does not recognise. Page IDs and URLs in their settings are switched.
- Links stored inside draw.io or Gliffy diagram files.
- Links inside comments (comments are not copied).
- Links to whiteboards, databases and embeds. They keep pointing at the source space.
- Links to an old version of a page, or to its history, keep pointing at the source page, because page history is not copied. The report lists them.
- Short links (
/wiki/x/…) are switched using Confluence's documented format. If one opens the wrong page, please tell us. - Text searches in CQL (
text ~ "…"ortitle ~ "…") are left exactly as written on purpose, even when they contain a space key orspace = …, because a search phrase cannot be told apart from a key. - Links that use the old title of a page renamed after the link was made (listed in the report).
- Space shortcuts in the sidebar, and links in the space logo or description.
- Links inside Jira issue macros, and Jira's links back to Confluence.
- Links from other spaces into the source space. They are not changed, because the source space still exists.
Not supported
- Confluence Data Center and Server.
- Copying to another Confluence site. Atlassian has its own tools for moving content between sites.
- Keeping a copy in sync with its source after the copy.