Journal file analysis/visualization

Journal files can provide key insights into database write activity for a number of purposes, ranging from debugging and performance analysis to incident investigation and response. I even once used the history of sets/kills of ^oddDEF to recover a week's worth of work that an intern accidentally reverted.

The existing management portal tools for searching through journals are very slow, because the APIs in %SYS.Journal.File/etc. operate directly against the files themselves. This makes log correlation difficult, best done as a manual binary search based on journal offset/timestamp.

It would be great if one could drop a journal file into a web UI, see an immediate overview of major activity areas at the global and/or process level, quickly drill down to journal activity in a particular timeline, and maybe even select range of global set/kill activity by process and/or transaction to either repeat or reverse.

For a bonus or a project for another day, do all of this real-time on an async mirror and throw in ML to flag and alert on anomalous journal activity and unusual global set/kill volumes.