Running DcrAssembleCoadd generates metrics that are very useful for evaluating the quality of the fit and the coadd, but which are only reported in the log. These should be persisted with the coadds separately from the logs, in a way that is easy to inspect by a user. Ideally these metrics would be available through the butler as well.
Metrics that would be useful to include: the number of iterations of modeling, the final convergence (possibly for each warp), the overall convergence improvement.
Note that Jointcal also generates useful metrics that are persisted only in the logs, so a system that works for both could be beneficial.