Thanks for the tip! Yes, this is a worthy effort to bring logs into the 21st century, even if it is still shackled to many of the assumptions that make logs wretched.
The good, at a glance:
* ditches writing out to local disk (fuckkk disks)
* unique ids, inheritance
* context persistence
* bundles up data by task rather than spraying log lines wantonly
* structured data
* appears to support high cardinality in context fields?
* ditches writing out to local disk (fuckkk disks)
* unique ids, inheritance
* context persistence
* bundles up data by task rather than spraying log lines wantonly
* structured data
* appears to support high cardinality in context fields?
The bad:
* indexes
* schemas
* Java
* doesn't gather nearly enough context automatically or otherwise, should have *hundreds* of context keys
* lacks info to use as spans
* unclear to me what exactly are "tasks"
* query model best suited for known unknowns
* indexes
* schemas
* Java
* doesn't gather nearly enough context automatically or otherwise, should have *hundreds* of context keys
* lacks info to use as spans
* unclear to me what exactly are "tasks"
* query model best suited for known unknowns
