Discussion Journal timestamps in the future?

[lanzz]

I'm trying to measure latency in processing of journal events in my tool and I'm comparing the event timestamp against wall clock time, but I'm regularly seeing events with timestamps 1-2 seconds into the future, which throws my latency measuring logic off (generally I'm aiming to process new events within a second, so if an event is stamped 2 seconds in the future and I process it in 2.5 seconds I will not be able to measure that I'm under target latency).

I've tailed the log files manually and I've confirmed visually that they do indeed appear with timestamps later than wall time, it's not some weird artefact of my code. I've confirmed that the HUD clock rendered in-game is in sync with my computer's time and I have no idea where this discrepancy comes from. My only guess is that events are actually stamped with server time and either the server is desynced from true time, or my computer is.

Has anyone else encountered this? If anyone else is measuring processing latency, I'm interested in what best practices would be.

 

[idahosurge]

I am just wondering, since the journal file is local to your machine wouldn't the time stamp logged for an event be your machine time when the event was received?

So I am wondering if there is a delay between when the event happened in the game and when the game sent to event to the journal file to be logged.

 

[lanzz]

I would expect the times to be local, yes. The problem is not that there is a delay between the event happening and being saved to the journal, a delay wouldn't be surprising; the problem is that the event saved has a future timestamp, e.g. if the local machine clock is 12:00:00, the event saved has a timestamp of 12:00:02. I'm only suspecting timestamps might be server time because I can't think of a plausible reason why a local timestamp would be generated in the future.