Handle timeseries of all zeros from Virgo for omegascans

Virgo has recently started uploading data that's full of zeros when not in observing, which causes FloatingPointErrors for omegascans: https://ligo-caltech.sentry.io/issues/4447067763/events/e730a532fbac4e4f9bc225112175af1a/?project=1425216.

We should handle the FloatingPointError to get rid of these Sentry events.

Edit: we already handle FloatingPointErrors: https://git.ligo.org/emfollow/gwcelery/-/blob/main/gwcelery/tasks/detchar.py#L145, and we've been logging these exceptions to Sentry. I think noting the IndexErrors and ValueErrors is still valuable for diagnosing issues with the data stream, but perhaps we can move the FloatingPointError handling elsewhere to avoid polluting Sentry.