More detailed test protocol:
Zone 'double' with two masters, M1 running this patch, M2 running current stable. 15.000 Services in the zone.

script used:

time curl -k -s -u root:icinga -H 'Accept: application/json' -X POST 'https://M1:5665/v1/actions/schedule-downtime?type=Service' -d '{ "start_time": 1544613000, "end_time": 1754613000, "author": "icingaadmin", "comment": "ALL DOWN"}'
echo 
sleep 1
curl -k -s -u root:icinga -H 'Accept: application/json' -X GET 'https://M2:5665/v1/objects/downtimes' | jq

Cropped output (it's a lot of text)

time: 2:22.96

      "attrs": {
        "__name": "test0!test-666!2381432d-9190-415d-8a3f-78dda94c22e3",
        "active": true,
        "author": "icingaadmin",
        "comment": "ALL DOWN",
        "config_owner": "",
        "duration": 0,
        "end_time": 1754613000,
        "entry_time": 1547213512.082966,
        "fixed": true,
        "ha_mode": 0,

For comparison, without this patch scheduling those downtimes takes 12 minutes

Thanks. It is ugly in this case, but larger rewrites in this region are TBD.

paused is never evaluated for these runtime objects, therefore updating the authority for all objects in an instance isn't needed.

@Crunsher thanks! Running test with 29,000 services on our QA system. Been running for 10 min. under 2.10.2 already. Will check final times tomorrow and then test a master snapshot.

Before patch result is: real 40m48.616s

command used:
time curl -k -u downtime:jU5... -H 'Accept: application/json' -X POST 'https://localhost:5665/v1/actions/schedule-downtime?type=Service' -d '{ "start_time": 1544613000, "end_time": 1754613000, "author": "icingaadmin", "comment": "ALL DOWN"}'

BTW, same command with remove-downtime is real 0m29.085s

@Crunsher including your patch using build 2.10.2+164.g21a592a.2019.01.17+1.xenial-0 it now completes more than 10 times faster: real 3m26.083s

ref/IP/9235

This is awesome. >60k downtimes in 7 minutes instead of 1-2 hours. Thanks @Crunsher!