Distributed Systems Error Logging

["Walden H. Leverich" <WaldenL@xxxxxxxxxxxxxxx>]

All,
 
What are people doing about "logging" errors that occur in our brave new
world of distributed systems? These days many "jobs" don't have joblogs,
plus you may have errors that occur inside of longer running (or never
ending) jobs like web sites. In "native" code we can always send a
message to QSYSOPR (and hope it's monitored) but in the distributed
world that gets either significantly harder, or downright impossible
depending on what machine the error occurs. 
 
We've somewhat standardized our internal systems on the syslog standard
that's popular in the unix world, but I'm wondering what others are
using.
 
-Walden