Icinga server retirement

once all Priority A metrics are deployed and we're satisfied with the alert manager operation, retire the Nagios server using a generous retirement delay (say, 90 days shutdown).

standard retirement checklist:

1. announcement and run in parallel with prometheus to see if alerts match, minimally (see also #41713 (closed) for an audit)
2. nagios (no need to remove from nagios, cleanup work covered by additional tasks below
3. retire the host in fabric
4. remove from LDAP with ldapvi
5. power-grep
6. remove from tor-passwords
7. remove from DNSwl
8. remove from docs
9. remove from racks
10. remove from reverse DNS

additional tasks:

• retire tor-nagios repository
• cleanup NRPE checks from Puppet
• remove tor-nagios package from APT repository