Is there some way of ensuring that various non-admin user's changes get versioned when they don't have access to the Commit button in the UI (from what I can tell that's only visible to Admin users)?

Something like every save creates a new version?

Or perhaps calling the commit endpoint on a schedule (with Admin credentials)?

If the second approach is the best option, is there some sort of limit on the number of commits and does the history need to be pruned over time?

This is for a self-hosted installation.

Just wanted to say thanks @admin, moving to Mongo brought down the folder cloning from upwards of 2 mins down to 11secs! UI and API seems much snappier as well.

(Note: Am using MongoDB Atlas Serverless, seems to work fine so far.)

One thing I found was that using a US database from the UK for development was crazy slow, presumably my internet connection or latency or something, but when I connected from the Docker image hosted in the same region, it was all good.

Next up, migrating the rendered reports. :)

Ok, thanks @admin!

Will I need to ensure the actual files (e.g. in the Mongo fs.files collection) have the same _id values, or are they just referenced against their filename property?

e.g. is the relationship between the reports collection and the fs.files collection between:
reports.blobName == fs.files.filename

Or does it use the ID in the filename or something, e.g. reports.blobName: reports/pdf/myTemplate/65a118152302904f0a473298.pdf => fs.files._id: 65a118152302904f0a473298?

I am currently using fs for my blobStorage provider, but am in the midst of migrating to MongoDB (GridFS).

Is there some built in method of exporting/importing rendered reports in the same way that we can migrate the templates? Or should I just script it instead?

I can get the list of reports with the /odata/reports endpoint in the old server, so I guess it shouldn't be too hard to script, just wanted to see if there was a built in way first.

Ah makes sense, thanks very much @admin.

Am in the midst of migrating to Mongo now, so hopefully that'll help matters!

Is there a more efficient way to add a user to a group, than issuing a PATCH request passing the entire updated list of user shortIds to /odata/usersGroups?

Current code (using got package for HTTP calls):

const usersGroupUpdateUrl = `${jsReportSettings.url}/odata/usersGroups(${usersGroupId})`;

// Send a PATCH request with full list of users
const usersGroupUpdateResponse = await got.patch(usersGroupUpdateUrl, {
  ...jsReportReqOptions,
  json: {
    users: [...usersGroupUsers, { shortid: newUser.shortid }],
  },
});

Can I POST just the new user payload { shortid: newUser.shortid } to that endpoint or something instead perhaps?

What I have works fine, but just means I need to query for the current list of users first, then update the array and then issue the PATCH request. Would be nicer to just be able to send the new user payload and have the backend handle all that.

I have a default set of templates in a folder structure, that I want duplicated for each new user.

I'm currently creating a new folder and then using the /studio/hierarchyMove endpoint to "clone" from the source folder into the new folder, e.g.

// Do the clone - from the existing folder to the newly created folder
const cloneUrl = `${jsReportSettings.url}/studio/hierarchyMove`;
const clonePayload = {
  source: {
	id: jsReportSettings.cloneFolderId,
	entitySet: "folders",
	onlyChildren: true,
  },
  target: {
	shortid: newFolderResult.shortid,
	updateReferences: true,
  },
  copy: true,
  replace: false,
};

(Not sure if there's a better approach, I got the above from examining the network calls when cloning a folder in the Studio...)

This does work, but I have an issue where the cloning process as outlined above is taking longer and longer with each new user.

Any ideas? Is there a better way to "clone" a folder? I couldn't find any documentation on the /studio/hierarchyMove endpoint, so not sure what the parameters do and if there's an option there that might help - updateReferences maybe?

Also, I'm currently using the file system - default jsreport-fs-store - for this (running a Docker image on Azure Container Apps with a "Premium" performance storage account - SSD backed).

Would switching to using a database backed store be faster, e.g. MongoDB or Postgres? Are there any benchmarks around what's fastest?