I've been having a mare of a time trying to work this out. I have Gigabit fibre through Slingshot (udm-se shows 2degrees now) and 99% of the time see as advertised performance, however I've recently upgraded my old asus rt-ac3200 for a udm-se so have a bit more eyes on monitoring and I've found an issue. I can't chalk this up to the hardware swap as I rarely use the service so have no idea of accurate prior performance. I've tested this on two separate pc's, both are hardwired with cat6 to the udm-se, one has an intel 2.5Gb port, the other has both 1Gb intel and 2.5Gb Asus, both have been tried. ONT and Router have been powercycled a few times. The udm-se wan is configured correctly as far as I am aware, ipv4 dhcp ipv6 disabled. v-lan 10, UPnP on (if that matters), I have a public ipv4 address and can connect to the internet fine.
I am running a pihole on the network but changing between the ISP and major DNS providers makes no difference. Nor does disabling it completely and having the devices point to the ISP recommended ones. Turning off IDS/IPS on the udm-se makes no difference either.
When I try to download a game from epic games launcher, my ping increases very quickly then I start getting packet loss warnings (as well as ping) from the udm-se. At first I noticed this in router management, then testing it during game download in cmd shows 150ms+ to 2degrees.nz as well as packet loss. The whole time, my actual download speed in epic games launcher shows as between 10-20MB/s with the occasional spike but predominately sitting at 10-12MB/s. The throughput on the udm-se fluctuates between 100-300Mbps while the ping and packet loss sky rockets and my network bogs down.
If I run a ping test during a steam download or p2p torrent client, whilst almost saturating my line with 80-100MB/s down, the ping is fine, I've only seen this with epic. Another odd factor is that if I cancel the download and restart many times, I'll occasionally get a connection that doesn't smash the ping, and will top out at 100+ MB/s but this is rare and doesn't coincide with peak/off peak times at all.
Running tracert to the epic download server (pulled from perfmon) with the download running at the time of writing this gives the following,
Tracing route to a23-216-54-139.deploy.static.akamaitechnologies.com [23.216.54.139]
over a maximum of 30 hops:
1 <1 ms <1 ms <1 ms unifi [192.168.0.1]
2 1 ms 2 ms 2 ms default-rdns.vocus.co.nz [101.98.0.119]
3 * 147 ms 143 ms default-rdns.vocus.co.nz [101.98.5.162]
4 78 ms 73 ms 71 ms default-rdns.vocus.co.nz [101.98.5.161]
5 2871 ms * * ae9.ngate-akl2.netarch.akamai.com [104.74.128.29]
6 103 ms 74 ms 71 ms a23-216-54-139.deploy.static.akamaitechnologies.com [23.216.54.139]
Running the same tracert with no download active gives,
1 <1 ms <1 ms <1 ms unifi [192.168.0.1]
2 1 ms 1 ms 1 ms default-rdns.vocus.co.nz [101.98.0.119]
3 5 ms 14 ms 6 ms default-rdns.vocus.co.nz [101.98.5.162]
4 4 ms 4 ms 4 ms default-rdns.vocus.co.nz [101.98.5.161]
5 6 ms 9 ms 17 ms ae9.ngate-akl2.netarch.akamai.com [104.74.128.29]
6 3 ms 3 ms 3 ms a23-216-54-139.deploy.static.akamaitechnologies.com [23.216.54.139]
I've not yet contacted Slingshot yet as I want to ensure it isn't at my end, anyone have any ideas?
