[Feature Request] Flag to log run info on removal from db

[HuntTheSun]

Would anyone be interested in a flag (could be handled by an env var also if you prefer) that logs run data to a json before removing the run from the db?

robotdashboard -r 'index=1:3' --log-removed-runs "/var/log/robotdashboard/removed_runs.json"

cat removed_runs.json | jq .
{
  "run_start": "2025-03-13 00:26:36.304669-05:00",
  "full_name": "Other",
  "name": "Other",
  "total": 114,
  "passed": 104,
  "failed": 10,
  "skipped": 0,
  "elapsed_s": "14.655",
  "start_time": "2025-03-13 00:26:36.305700",
  "tags": "prod,project_1",
  "run_alias": "output-20250313-002636",
  "path": "D:\\CodeProjects\\robotframework-dashboard\\.\\tests\\robot\\resources\\outputs\\output-20250313-002636.xml",
  "metadata": "[]",
  "project_version": "2.3"
}
...

# You could then for example use the log for cleaning up leftover log.html files by getting the corresponding run_alias
cat removed_runs.json | jq .run_alias
"output-20250313-002134"
"output-20250313-002151"
"output-20250313-002222"

Would that be something that could be of use to the community or would it not be worth the extra flag in your opinion?
I implemented it roughly to see check much hassle it would be and its not really a hassle or hacky (for my tastes).
I would of course volunteer to implement it if you like the feature.

Thanks in advance!

[timdegroot1996]

What would be the end goal of having this data? I understand keeping some form of it available, but what usecases do you see?

Apart from maybe the only usecase I can think of which is reinserting it if you accidentially deleted it...

[HuntTheSun]

Hi,

for example my team would like to have an easy way to identify unneeded .xml and report/log.htmls of which the runs have been removed. We have a decently complicated retention schedule and pipeline, so it would be great if we could grab for example the aliases of the runs that were removed in order to identify which files to clean up.
It could also allow us to more easily script a sync functionality for another software that pertains to tests, so that I can more easily identify and remove runs that were removed from the dashboard in other software.

We probably only need the aliases for our use case, but I thought having a feature with all data would allow for more use cases I haven't thought of.

[vladislav1q]

Does rf-dashboard supports autoremove output.xml from database by timeout ?

[HuntTheSun]

Do you mean autoremove run data based on run_start "age"?
Or do you mean removing the output.xml from filesystem itself?

[vladislav1q]

removing the output.xml from filesystem, if upload time is N days before current_timestamp

[timdegroot1996]

I think automatically removing output.xml's is not in scope of robotdashboard. If a user wants to do that with a script of some kind they are free to do so, but the only actual removal of data I want to do with the tool is in the database.

So for me age would be a good flag (similar to limit but then based on how old they are, like 1 month or something, or 30 days, format will have to be discussed further).

In addition to this I think it's nice to create what's described here to write the deleted info to a json file or some other format which can then be used for other stuff like you mentioned, maybe even deleting the output file if users want to do that themselves.

[vladislav1q]

I think, that json should contain delete timeout, that should be implemented by timer for execution of sql commands.

[HuntTheSun]

I think, that json should contain delete timeout, that should be implemented by timer for execution of sql commands

@vladislav1q
Do you mean the db table "runs" should contain a column "scheduled_deletion_timestamp" or something like that, that could be passed on adding the run?
I think that adds a decent amount of scope and headaches for something that could be achieved via the existing mechanisms.

For example, if you know a run should be deleted in the year 2026 on calendar week 42, you could just give it the tag del_2026_week_42 and weekly run a scheduled script that does it like this

year=$(date +%G) # gets year
week=$(date +%V) # gets calendar week
deletion_tag="del_${year}_${week}"

robotdashboard -r "tag=${deletion_tag}"

If you want to discuss the feature of a scheduled deletion time, could you please open a dedicated Issue for it?
I would prefer keeping the discussion in this Issue focused on its topic if that is alright with you.

[HuntTheSun]

In addition to this I think it's nice to create what's described here to write the deleted info to a json file or some other format which can then be used for other stuff like you mentioned, maybe even deleting the output file if users want to do that themselves.

Should I look into outputting into other formats also or do you have any other suggestions about the usage or subfeatures?
Otherwise I would go ahead and submit a PR if thats okay.