Just a suggestion. IMO it is way more important to see at what hour you acquired specific users compared to when they dropped off. Especially if you are running certain ads for example etc. We could also add another report "first" and "last action" hour but I think that not both are actually needed and keeping it simple be better.
Performance wise this shouldn't have really any impact as the index be still used in the back by the where statement.
Could also save it for Matomo 4 or so.
It should be fine. The only "problem" could be that, when you sum the visits for this report Visits by server time, then it won't necessarily match the overall visits for the day, because if a visit overlaps two days they will appear for DAY1 in the Visits by server time graph, but won't be included in the DAY1 overall visits metrics. (To process overall visits we use the server time at the end of the visit.)
that said I don't think it's a blocker, but we may get a few questions, it should be fine :+1:
By default we create a new visit after midnight correct? So it might reduce the problem?
Good point, forgot about this, was done in #6217, so it's all good
Fixed ui tests. Looked at the report description and looks like the report description was already written generic and not that it was based on the last action time. But maybe we should make it a bit more clear on what it is based?
GetVisitInformationPerServerTime report class still uses the
VisitLastActionTime dimension. Should this change?
Noticed the GetVisitInformationPerServerTime report class still uses the VisitLastActionTime dimension. Should this change?
I'm actually not sure. Might be a good idea. I could change it to first action time dimension and we see what happens re tests. Shouldn't affect too much I suppose. Possibly processed reports may be better.
just pushed the change now.