Timeseries data for infrequent data

Looking for advice on first time dive into Mongo’s time series data.

We have several thousand IOT devices which connect to our back end servers periodically throughout the day. I would like to use fact that IOT device has connected to back end server to post entry into time series collection to create a historic log of connectivity - nothing fancy.

The IOT devices will connect approximately every 15 minutes and seeing as I don’t need this granularity and just require knowing that a device connected at least once a day all I need is one entry per day.

One way to do this would be to check if entry already exists in db when IOT device connects and simply not add another entry if time period is same day - but this means a read per IOT per 15 minute interval - again not too heavy.

Or, I could insert timestamp into memory based redis database (running local to the server) and query this redis table for entry and if not finding it then inserting into mongo - should be less resource intensive and I already have redis available on the server.

Or is there another way where I could do a db write and it updates that specific days entry - still a database access per IOT query. When I first saw word 'granularity` I thought maybe I could leverage this - wishful, I know.