macOS Server 5.3 Calendar pg_ctl not starting - macos

Question

After updating macOS Server to 5.3 (running on macOS 10.12.4) my Calendar & Contacts have stopped syncing.
It seems that it's having trouble starting Postgres for cluster /Library/Server/Calendar and Contacts/Data/Database.xpg/cluster.pg and possibly trouble with the agent too.
The GUI seems to think that the Calendar & Contacts services have started and are available, but when I run $ sudo serveradmin fullstatus calendar from the command line I get:
calendar:setStateVersion = 1
calendar:readWriteSettingsVersion = 1
calendar:state = "STARTING"
calendar:contactsState = "STARTING"
calendar:calendarState = "STARTING"
System log is being spammed with:
Apr 22 11:58:42 com.apple.xpc.launchd[1] (org.calendarserver.agent[44649]): Service exited with abnormal code: 1
Apr 22 11:58:42 com.apple.xpc.launchd[1] (org.calendarserver.agent): Service only ran for 0 seconds. Pushing respawn out by 10 seconds.
Apr 22 11:58:52 com.apple.xpc.launchd[1] (org.calendarserver.agent[44659]): Service exited with abnormal code: 1
Apr 22 11:58:52 com.apple.xpc.launchd[1] (org.calendarserver.agent): Service only ran for 0 seconds. Pushing respawn out by 10 seconds.
Apr 22 11:59:02 com.apple.xpc.launchd[1] (org.calendarserver.agent[44668]): Service exited with abnormal code: 1
Apr 22 11:59:02 com.apple.xpc.launchd[1] (org.calendarserver.agent): Service only ran for 0 seconds. Pushing respawn out by 10 seconds.
Apr 22 11:59:07 com.apple.xpc.launchd[1] (org.calendarserver.calendarserver[44676]): Service exited with abnormal code: 1
Apr 22 11:59:07 com.apple.xpc.launchd[1] (org.calendarserver.calendarserver): Service only ran for 0 seconds. Pushing respawn out by 60 seconds.
Here's the output of $ sudo /Applications/Server.app/Contents/ServerRoot/usr/sbin/calendarserver_diagnose
Any ideas?
OS Build: 16E195
Server Build: 16S4123
/Library/Server/Preferences/Calendar.plist exists and can be parsed
Prefs plist says ServerRoot directory is: /Library/Server/Calendar and Contacts
ServerRoot volume ok
/Library/Server/Calendar and Contacts/Config/caldavd-system.plist exists and can be parsed
/Library/Server/Calendar and Contacts/Config/caldavd-user.plist does not exist
Configuration:
Calendar and Contacts service processes:
USER PID %CPU %MEM RSS ELAPSED STARTED COMMAND
root 42554 0.0 0.1 11072 07:49 Sat 22 Apr 11:32:16 2017 servermgr_calendar
Serverd status:
org.calendarserver.agent is enabled
org.calendarserver.calendarserver is enabled
org.calendarserver.relocate is enabled
Disk space on boot volume:
Filesystem Size Used Avail Capacity iused ifree %iused Mounted on
/dev/disk1 999G 777G 222G 78% 8520180 4286447099 0% /
Disk space on service data volume:
Filesystem Size Used Avail Capacity iused ifree %iused Mounted on
/dev/disk1 999G 777G 222G 78% 8520180 4286447099 0% /
Disk space used by Calendar and Contacts service:
20K /Library/Server/Calendar and Contacts/Config
1014M /Library/Server/Calendar and Contacts/Data
200M /Library/Server/Calendar and Contacts/Logs
Postgres status for cluster /Library/Server/Calendar and Contacts/Data/Database.xpg/cluster.pg:
pg_ctl: no server running
Agent:
Attempting to send a request to the agent...
Can't connect to agent: timed out
Server connection:
Traceback (most recent call last):
File "/Applications/Server.app/Contents/ServerRoot/usr/sbin/calendarserver_diagnose", line 14, in <module>
load_entry_point('CalendarServer==9.1a1.dev0+56b4197875debefef19d9c19840f903a8e480c88.head', 'console_scripts', 'calendarserver_diagnose')()
File "/Applications/Server.app/Contents/ServerRoot/Library/CalendarServer/lib/python2.7/site-packages/calendarserver/tools/diagnose.py", line 145, in main
connectToCaldavd(keys)
File "/Applications/Server.app/Contents/ServerRoot/Library/CalendarServer/lib/python2.7/site-packages/calendarserver/tools/diagnose.py", line 584, in connectToCaldavd
url = "https://{host}/principals/".format(host=keys["ServerHostName"])
KeyError: 'ServerHostName'