Fix bug 1653 to update osm_health increasing time_for_failure

Recent tests showed that the 5 minutes assigned to the time for checking
failures in pods were not enough. This change increases the time to 7
minutes, which will be enough even for the long tail cases.

Change-Id: Ic9c721811a187292c1da6f421f2ed4311b7f4622
Signed-off-by: garciadeblas <gerardo.garciadeblas@telefonica.com>