Collection validation has two classifications for its results: warnings and errors. Typically warnings do not indicate data corruption and errors do indicate data corruption. However, currently there are some validation results which we report as errors despite not indicating real data corruption. A couple examples of this are document schema violations and unrecognized index spec fields. We should consider reclassifying some of these so that they are not reported as data corruption. However, we should make sure we can still catch these cases if they occur in our testing.