Changing a stock count with ExecuteUpdateAsync and recording the reason with SaveChangesAsync looks like one job, but those are two separate transactions unless you say otherwise. If the second fails, the count has changed and nothing explains why. Here's that failure, and the fix.
EnableRetryOnFailure takes a list of numbers, and that's all it takes. With MySQL I needed two more things: to tell a dropped connection from a full pool, which share one number, and to close a broken connection before retrying. A class of your own does both, in about 20 lines.
EnableRetryOnFailure looks like a switch, and with Oracle's MySQL provider, it isn't one: until you list error numbers, it retries nothing. Here's the list I use, how to confirm it's working, and how long a request can wait.
As soon as a DbContext has a retrying execution strategy, a plain BeginTransactionAsync stops working. The strategy has to know the whole group of statements, so it can run them all again. It only takes a wrapper, with one detail in the lambda and one at the top of it.
When a transaction doesn't do what you expected, the first question is whether it began and committed at all. EF Core logs both, at Debug level, under a category of their own. One line in your settings shows them, without turning on everything else.
SaveChangesAsync protects one call. When several calls have to succeed or fail together, you need a transaction of your own, and it takes three lines. The last of them decides everything: without it, nothing you did is saved.
Yes. Each call to SaveChangesAsync runs in a transaction of its own, so if one row fails, none of them are saved. That's easy to prove with MySQL's own log, and it also shows you where the protection stops.
Cancelling an order puts its stock back. Cancel it twice, and the stock goes back twice. Reading the status and then updating leaves a gap for exactly that to happen. Put the expected status in the WHERE, and only the first update can change the row.
To see locks, you need rights that an application's account shouldn't have. So don't give them to it. Make a small second account just for looking: it can read who's waiting for whom, and it can't read or change your tables.
Reading a quantity, checking it, and then writing a new one takes three steps, and two people can both pass the check. A single UPDATE with the check in its WHERE takes one step. MySQL holds the row's lock from the check to the change, so there's no gap to fall into. It's the simplest safe way to take stock.
