We don't tend to have scaling issues with the DB because we have an HA database system. I don't know the details, but there are several "satellite" servers that accept queriers, and together they form a sort of "logical database" that is relplicated, in turn, to a more solid archive. I could be explaining it wrong, as I didn't set it up.
The local daemon dying *is* one of the issues. Until recently, apps dealt with this in undefined ways (by which I mean the authors chose, there was no standard). Just recently, we decided that we'd write app-named files to a specific directory, which the daemon scans and uploads (and cleans) at startup. I have reservations about this, though -- it seems like asking for trouble.
Powers that be know there are issues with this too, but no one (including me) has come up with a better idea yet...
In reply to Re^4: Central logging methods and thoughts
by radiantmatrix
in thread Central logging methods and thoughts
by bwelch
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |