In some email I received from Rainer Gerhards, sie wrote:
> Hi all,
>
> A current discussion on the loganalysis mailing list has shown that many
> people are not quite happy with the syslog timestamp. Specifically, the
> absence of time zone information is causing problems to many
> geographically spread locations.
>
> The problem with the current [RFC3164] timestamp is that it is not
> possible to reliably consolidate log data over multiple time zones, not
> even when proper time synchronization is in place. The only way to do
> this would be to configure the central collector(s) to know each device
> generating syslog messages and be configured to adjust the reported time
> accordingly. Obviously, this is error-prone and work-intensive.
>
> As such, I am asking for a modification to the current syslog-sign
> draft, specifically section 2.2 (HEADER) where the TIMESTAMP is
> discussed. I suggest that a timestamp as described in [RFC3339]
> "date-time" format should become recommended but the [RFC3164] timestamp
> should still be allowed (but be depraciated). That way, it would be
> possible for a collector to detect both formats and act on the less
> precise accordingly.
>
> Any support for this on the WG?

I think that this is getting into message content and that's outside
the charter for this WG.


Reply via email to