You have some complexity there that requires some context and knowledge of the application over time. It would likely take some api calls and some logic to look for deltas over time. Like you want new transactions. How far back do you have to go to look for any instance of them…
For transaction growth you can maybe take a more simple approach to first identify if you have growth and then make some queries to find the new ones.
This will show you distinct transaction names over time and compare them with yesterday.
select uniqueCount(name) from Transaction since 23 hours ago TIMESERIES 1 hour COMPARE WITH 1 day ago
Typically, it’s not so simply to do this since applications with a number of transaction have some variable throughput of them over time. Like there may not be any traffic on a transaction at 3am.
You can try to table the transactions and then sort and look for gaps on an hourly basis if you don’t want to write some code for the logic.
SELECT filter(count(*), where hourOf(timestamp) = '12:00') as '.12',filter(count(*), where hourOf(timestamp) = '13:00') as '.1pm', filter(count(*), where hourOf(timestamp) = '14:00') as '.2pm' from Transaction facet name since 23 hours ago limit MAX
Its possible someone else has some clever tricks.