Gaps reported for epochs,tasks which already have an OK

[placer14]

After doing some exploration in our local DB instance, we see many cases where a gap find produces a GAP report for epoch,tasks already have OK processing records. (Specifically, the OK existed and still produced a GAP report after the fact.) Notably, all of these observed rows also contain rows with SKIPs, which may be related.

Here are some records I found in our production DB to illustrate (query at the end): https://gist.github.com/placer14/df45f16be350ae7c3223584fe10a6039