google-analytics-measurement-protocol | Python package to send hits | Analytics library

We have an online marketplace website with auctions. We would like to measure won auctions in Google Analytics, but this is not a frontend event, but a backend event (a timer for auction end expires). The user, who leads with the highest bid placed, wins the auction, once the timer runs out. The auction can last several days.

We would like to know where the winners of the auctions came from to place their first bid. So we are interested in source, medium and campaign. Currently we store in our backend DB a clientId and if present in URL also source, medium and campaign (or gclid). Later a backend job sends the data of finished auctions using measurement protocol to Google Analytics. But using this way we gather too few source, medium, campaign data in comparison to all clientIds for user bid events. User can come from search or ad to the website, click through the website and then place a bid and information about source, medium or campaign is lost.

So my question is

how to gather more source, medium and campaign data at the time when a bid is placed? Data will be send to GA using measurement protocol a few days later.

I did some research:

Should we use non-interaction hit? I found the suggestion here. But it is only briefly explained. Would the GA pair correctly automatically a frontend bid event with a few days later sent hit using MP? We know clientId.

Will it help us to define a new dimension for clientId that can be used to fetch source, medium and campaign data using report api v4 when a bid is placed using the steps described here?