index cleanup fixes while applying retention

[sandeepsukhani]

What this PR does / why we need it:
This PR includes two fixes:

1. Add a metric name if missing in labels used for calculating index entries for the series clean up since it is used while building the series IDs. The test did not catch this because the test data included the metric name label. Without the metric name, wrong series IDs were calculated, which caused labels entries not to get cleaned up.
2. While calculating table interval from table name, we calculate the start time first and then add 24h to it. The problem with this is it would end up covering the start time of next days table. This is fixed by subtracting a millisecond from the end time of the interval after adding 24h to the start time. There was no harm caused by the overlap with the next day, I am just fixing it to avoid index entries being calculated unnecessarily for the next day as well.

Checklist

• Tests updated
• Add an entry in the CHANGELOG.md about the changes.

[cyriltovena]

LGTM

Would be great to have regression tests.