>>> Has anybody used this approach of running a VFP application in the Windows Task on a Server? Any potential issues or things that I need to be aware of?
>
>Works well, but be careful about proper app close down. E.g. if terminating after error, tablerevert() open Views so they don't prevent shutdown. Also note that Task Scheduler allows you to terminate an instance if it's still running, which can damage local data if your app is stalled or lose records if still processing.
>
>Finally, consider Heartbeat and Raisealert logs in addition to any application logging. A Heatbeat log confirms the app started fine and has something useful to process: an interface engine can monitor the timestamp and alert IT staff via txt message or email if the heartbeat is delayed or absent. A Raisealert log means a hard stop happened (e.g. can't access database or server share) which again can be monitored to raise an alert if it's ever updated. This is better than expecting support people to check app logs several times a day or waiting for end users to notice there's no output. Writing local logs is more reliable and risks less support burden than (for example) a feature to send a failure alert email, which itself can fail.

Thank you very much for your suggestions.