I'm having constant packet loss to my upstream gateway (102.182.203.1) and other services. Some days I'll have 100% packet loss for a few minutes multiple times a day. Google Meet will drop out at least 50% of the time, making it very hard to talk to my colleagues.
This is the last 1 hour (2023-11-30 7am to 8am)
View attachment 1626703
And over the last 24 hours (2023-11-29 to 2023-11-30).
View attachment 1626705
Note the packet loss is higher than the graph represents. This is because Grafana is averaging the values over a larger period as you increase the window. So instead of averaging intervals over a 5s period, it averages over a 30s period, when your didn't have packet loss for say 20s, so it shows the average of say 20% and 0% which is 10%.
For example if zoom into the gateway graph around midnight where it shows < 20% packet loss, the loss is actually MUCH higher, up to 100% (Date for below is 2023-11-29 going into 2023-11-30)
View attachment 1626709
So there is definitely a problem with my line or something just upstream of it.
Then the other problem is further upstream, I guess with Afrihost or their peers (Date the below is 2023-11-29)
View attachment 1626715
Google DNS (8.8.8.8) and Showmax NAPAfrica (169.255.195.23) experience the same packet loss at the same time, while AWS/Prime (52.85.24.6) has no issue. I assume this could be because AWS is hosted in Cape Town and Showmax is in JHB?
To show it isn't an isolated incident, here are a few more graphs. I've included the date ranges in the screenshots (top left). Again, if a view is zoomed out the values will be lower because of grafana averaging, not because they are actually lower.
View attachment 1626723
View attachment 1626725
View attachment 1626727
Edit: I am on FrogFoot in the Cape Town region.