The way we typically do this at our site is to have a "Start Time" column, an "End Time" column, and a timeout flag.

The reason for the start/end times is that you can subtract start from end and find out how long the poll took (usually in milliseconds, but for some systems in millionths of a second).

The timeout flag tells you that the transaction ended because of a timeout.

The reason you want to do this is that your timeout value might change over time ( 30 seconds this month, 25 last month, 45 two months ago), or you might expand the system to multiple customers where each has his own timeout.

Reporting timeouts for a period of time (1 day, 1 week, 1 month) is now much easier, because you can use the start timestamp as a key value field in the time period (be sure to create indexes on start and end time).

Just my thoughts.