I was interrupted while writing this, and connectivity has been restored before I got back, so now just posting for posterity.

It looks like Spark started blocking, in some form, OVH traffic in the US around 1AM. OVH traffic from AU doesn't seem to be affected, e.g. the SoYouStart website. Perhaps Spark has null-routed OVH (and other network/s) due to an attack.

From OVH GRA (France): 

traceroute to spark.co.nz (146.171.248.36), 30 hops max, 60 byte packets
 
 1  51.255.x.x (51.255.x.x)  0.431 ms  2.677 ms  0.693 ms
 
 2  10.17.132.114 (10.17.132.114)  0.554 ms  0.585 ms 10.17.132.106 (10.17.132.106)  0.685 ms
 
 3  10.73.0.138 (10.73.0.138)  0.259 ms  0.242 ms 10.73.0.58 (10.73.0.58)  2.547 ms
 
 4  10.95.33.8 (10.95.33.8)  0.484 ms 10.95.33.10 (10.95.33.10)  1.879 ms 10.95.33.8 (10.95.33.8)  1.060 ms
 
 5  lon-thw-sbb1-nc5.uk.eu (91.121.215.179)  3.704 ms  3.677 ms be102.lon-drch-sbb1-nc5.uk.eu (213.251.128.65)  3.680 ms
 
 6  * nyc-ny1-sbb1-8k.nj.us (192.99.146.127)  80.585 ms *
 
 7  nyc-ny1-sbb1-8k.nj.us (192.99.146.127)  76.311 ms be102.bhs-g2-nc5.qc.ca (192.99.146.138)  78.995 ms be102.bhs-g1-nc5.qc.ca (198.27.73.204)  86.946 ms
 
 8  be102.bhs-g1-nc5.qc.ca (198.27.73.204)  87.750 ms be101.chi-ch2-sbb2-8k.il.us (192.99.146.141)  94.137 ms be102.bhs-g1-nc5.qc.ca (198.27.73.204)  83.024 ms
 
 9  be101.chi-ch2-sbb1-8k.il.us (198.27.73.207)  98.901 ms  98.798 ms *
 
10  10.200.3.195 (10.200.3.195)  146.758 ms *  150.551 ms
 
11  f0-0.pabr1.netgate.net.nz (198.32.176.97)  145.714 ms  149.214 ms 10.200.3.195 (10.200.3.195)  147.476 ms
 
12  * * *
 
[...]
 
30  * * *

From OVH SYD (Australia): 

traceroute to spark.co.nz (146.171.248.36), 30 hops max, 60 byte packets
 
 1  * * *
 
 2  10.161.33.125 (10.161.33.125)  0.179 ms 10.161.33.124 (10.161.33.124)  0.199 ms  0.169 ms
 
 3  10.133.20.24 (10.133.20.24)  0.614 ms 10.133.20.26 (10.133.20.26)  0.370 ms 10.133.20.20 (10.133.20.20)  0.609 ms
 
 4  10.75.8.30 (10.75.8.30)  0.107 ms  0.128 ms 10.75.8.32 (10.75.8.32)  0.170 ms
 
 5  172.20.136.2 (172.20.136.2)  3.111 ms 172.20.136.0 (172.20.136.0)  1.244 ms  1.222 ms
 
 6  72.251.0.9 (72.251.0.9)  2.367 ms  2.322 ms 72.251.0.11 (72.251.0.11)  3.030 ms
 
 7  10.200.6.1 (10.200.6.1)  0.502 ms  0.575 ms 10.200.6.9 (10.200.6.9)  0.496 ms
 
 8  4648.syd.equinix.com (45.127.172.168)  0.583 ms  0.547 ms  0.311 ms
 
 9  et10-0-0.tkbr12.global-gateway.net.nz (122.56.127.49)  22.999 ms et11-0-0.akbr7.global-gateway.net.nz (122.56.119.29)  23.584 ms et5-0-2.akbr7.global-gateway.net.nz (122.56.119.25)  23.911 ms
 
10  ae10-10.akbr7.global-gateway.net.nz (202.50.232.30)  24.990 ms 122.56.119.222 (122.56.119.222)  23.784 ms  23.766 ms
 
11  122.56.119.222 (122.56.119.222)  24.699 ms * *
 
12  * * *
 
[...]
 
30  * * *

Note that the traceroute from AU reaches NZ, but the one from FR stops in the US.

Just tested it and it's working again - phew! Thanks to whoever restored it 🙂
@cyril7 any idea what happened?

I am on 2 degrees and its no dice. 10:52 pm Sunday.

I have been having streaming issues on 2 degrees for some time so not surprised.

The answer is this was planned work carried out. I have no idea if/what the front line teams get told about work like this though should they get any queries.