[Needs clarification] Uploading report as html.gz using add-log-file is successful but throws error

[nadhor2001-blip]

Hello,
to upload logs and reports to dashboard I am using one and the same (self-created) function. (info: postfix is unique for every run, so log, output and report have always the same postfix when being uploaded)
In it I first use gzip.compress to compress the html file and then post(dashboard_url/add-log-file, compressed_file) to upload the file.
With log_postfix.html this works without issue. But when trying to upload report_postfix.html the dashboard server says:

ERROR: something went wrong while adding the log file or updating the database: A problem occured while adding the log file, check the console message!
ERROR: There was no output with the name report_postfix.xml or report_postfix.html in any of the existing outputs in the database. 

The same happens, when uploading the pure .html file with the swagger UI.
But the upload worked, and I can open the report using the dashboard.
Is there another - recommended - way, to upload reports? In the documentation I can find, that it is possible to use uploaded reports, but no way on how to do it specifically.
Regards,

[timdegroot1996]

Hi @nadhor2001-blip, thanks for the detailed report, clarification below.

What's happening

The /add-log-file endpoint isn't really a generic file uploader — it's specifically for linking a log.html to a stored run. After writing the file to disk it tries to match it back to a run in the database, and that matching step expects a log-named file:

• It derives the expected output filename by replacing log → output and .html → .xml in the uploaded filename, then looks for a run whose stored path matches.
• For report_postfix.html there's no log substring to replace, so it ends up searching for report_postfix.xml / report_postfix.html, finds no matching run, and returns the error you saw.

The file write happens before that matching step, which is why the report still lands on the server and opens correctly — but the endpoint reports failure because the DB link step found nothing to update. So for reports the error is currently expected, and (for now) harmless.

Recommended way to make reports work today

Reports are not tracked in the database. They're reached through the link that Robot Framework already builds into log.html itself (top-right corner). For that link to resolve, the report just needs to sit next to the log with a matching name:

1. Name the report the same as the log, but with log → report (e.g. log_postfix.html → report_postfix.html).
2. Upload it so it lands in the same robot_logs/ folder as the log (which your current flow already does).
3. Open the report from inside the log via that built-in top-right link.

So your existing upload is actually fine — you can ignore the error for the report file. Just make sure the names line up (log_<postfix> / report_<postfix>). This is documented here: Accessing Reports.

Possible small improvement

One thing we could improve on our side: when the uploaded filename contains report, the endpoint currently still tries to match it to a run and fails. Instead it could either skip the DB-matching step entirely for report files (assume the name is already correct and just save it), or match it to the corresponding log/output by replacing report → log/output like it does for logs. Either way it would stop returning a misleading error for a file that was actually saved fine. What do you think is a good way forward?

A couple of questions for you

1. Does the flow above (name the report report_<postfix>.html, upload to the same folder, open it via the link inside the log) actually work for your setup? Want to confirm it's a usable path for you before we treat it as the recommended approach. If it doesn't work we should change the docs so feedback is appreciated!
2. Was the documentation unclear on this? If so, what would have made it clearer — we'd like to update the docs so the way to upload and use reports are clearer in the docs, regardless of the upload error.

Thanks!

[nadhor2001-blip]

Hi,
thanks for your answer.
Concerning your questions:

1. The workflow currently works for me, I can access log, access report from log and access log from report. I do not think anything needs to be done concerning the db.
2. The documentation was quite clear. (else I wouldn't even have tried uploading the log using add-log-file)

Currently I see two possible improvements:

1. continue allowing add-log-file to be used to upload report_xxx_.html. But add a check for reporthtml. If matched, check if loghtml resides in same folder and if so, an info like "SUCCESS: the report file has been placed and is accessible from your log file." If no matching log*.html file is file, perhaps a "WARNING: the report file has been placed but no matching log file has been found. You may not be able to access your report from your log file."
2. add another access-point "add-report-file".

I personally would prefer solution 1, as it already works. With solution 1 one could also simple expand the documentation, mentioning that add-log-file can also be used to upload report*.log and it is strongly recommended to upload the report AFTER the log, to get no warning.
Thanks again for your work and regards.

[timdegroot1996]

Glad to hear that it at least isn't wrong in the docs!

I like your suggestion 1 the most as well. I think it makes sense to add some handling for it to not give an error but only a warning if no matching file is found. And then to clarify that some more update the docs to explain this handling.

Option 2 is one I also considered but rejected. The main reason being that currently the only way to "reach" the logs is by clicking graphs or the overview links, and then you can only reach the report through the log file. Since there's no direct link available it's kind of strange to facilitate an endpoint to upload them but make it quite hard to actually open them.

As I'm typing this I'm considering another direction (option 3) which would kind of combine your options 1 and 2. What do you think about:

1. Adding a separate enpdoint to add report files (practically a copy of how it already works for the add-log endpoints, but the difference being that it looks for matching outputs by replacing report for output to match and give warnings/errors if it's not found.
2. Adding a link on the overview page cards which links to report.html so that they can actually be reached. (This wouldn't break the already existing flow of opening log file > top right > open report, but it would allow a clean entry from the dashboard itself)
3. Adding some better docs that explain how to handle the report files.

I think my preference would be to completely split it with a clean entry point (option 3 I described above). This would make it the most user friendly and maintainable. Right now it's always a bit confusing because you're abusing the log endpoints...

Curious to hear what you think about this!

[nadhor2001-blip]

Hi.
adding another entry-point was what I meant with option 1 (access-point). This would make a lot more work, I think.
We would not only need: /add-report-file, but also /remove-report and /get-reports, to keep everything consistent.
Adding an additional link on the overview pages, I can't quite imagine. Some kind of bubble over each run? I do not think I'd like this. There is already so much Info on the dashboard, no need to add ever more (IMHO).
Perhaps instead, if really adding a new entry-point /add-report-file, add a parameter "link". If this parameter is True, a click on the dashboard links not to the log, but to the report. Meaning instead of putting ..../log_.html in Table runs, Column: path, .../report_.html would be saved.
This would require no change on already existing code and would just add another feature.
Regards,