We are running k8s cluster in AWS Singapure region. We also have self-managed EC2 NAT instance of type t3.xlarge
. According to spec, it has 1 Gbps baseline speed, and burst up to 5 Gbps.
When running speedtest on the instance, I'm seeing 1gbps for download and upload. However, real speed is very different when I'm trying to download a file from server in EU. Downloading a test file from hetzner (https://hel1-speed.hetzner.com) shows around 100mbps.
At the same time, Network in (bytes)
chart in AWS is showing 1gbps.
Why are we losing so much speed? I understand that there is latency between regions, but it shouldn't be that large? mtr is showing large packets loss:
13. if-ae-41-2.tcore2.svw-singapore.as6453.net 72.4% 30 200.5 202.4 199.4 215.0 5.2
14. if-ae-33-2.tcore2.svw-singapore.as6453.net 86.2% 30 180.5 180.7 180.4 181.2 0.4
15. (waiting for reply)
16. if-ae-36-7.tcore1.av3-toyohashi.as6453.net 65.5% 30 201.8 201.4 200.7 203.0 0.7
17. if-ae-2-2.tcore2.av3-toyohashi.as6453.net 72.4% 30 193.3 193.4 192.9 194.2 0.3
18. (waiting for reply)
19. if-bundle-34-2.qcore2.lvw-losangeles.as6453.net 73.3% 30 202.1 202.4 201.9 202.9 0.4
20. if-bundle-2-2.qcore1.lvw-losangeles.as6453.net 60.0% 30 199.2 199.2 198.8 200.1 0.4
21. ash-speed.hetzner.com 60.0% 30 253.0 255.0 252.6 268.5 4.7
I would appriciate any suggestions on where else I can look to solve this.