Peak hour high local latency

Lepep

Active Member
Joined
Apr 4, 2018
Messages
44
Reaction score
12
FNO: TTConnect
ISP: CISP
Location: Johannesburg
Router: Ubiquiti Cloud Gateway Ultra

I've had a weird issue for well over a year now which CISP and TTConnect support cannot figure out.
During peak hours from ~6-10PM I will, on MOST days, have ~100ms to ALL Johannesburg servers, never mind the higher latencies beyond JHB.
I normally have 0-5ms with 40ms jumps here and there until peak hours where it stays around 100ms before it falls back to normal, this is obviously a disturbance to my online gaming.

Hop tests always show the first hop after my router being 100ms which carries through to the endpoint, and I've even done ONT tests with 100ms on the initial hop while on a call with TTConnect telling me in that same moment that it looks good on their side.

Where is this latency coming from then if:
- The first hop already shows 100ms (Router hop stays at 0ms)
- This doesn't happen every night, but most nights around 6-10PM
- LAN was checked and found good, fibre was checked + changed to a less congested splitter and found good?
 
Hi there, can you post your support ref and an MTR for CoolEscalator to investigate?

That sounds like congestion/line related.

Potentially switch congestion.
 
I've had many tickets with you guys about this, but here is the latest:
#COOL-20260315-2623413

I've had a TTConnect technician clean my fibre and put me on another splitter with 24/64 clients, found no noise.
I am aware it is congestion of some kind, but I cannot keep going on this loop on doing ONT tests while CISP and TTConnect both saying it's not their issue.
 

Attachments

  • Screenshot 2026-03-15 202757.png
    Screenshot 2026-03-15 202757.png
    50.1 KB · Views: 26
  • Screenshot 2025-06-07 210135.png
    Screenshot 2025-06-07 210135.png
    76.4 KB · Views: 25
I've had many tickets with you guys about this, but here is the latest:
#COOL-20260315-2623413

I've had a TTConnect technician clean my fibre and put me on another splitter with 24/64 clients, found no noise.
I am aware it is congestion of some kind, but I cannot keep going on this loop on doing ONT tests while CISP and TTConnect both saying it's not their issue.
HI there, thank you for sharing your info.
Would you also mind enabling ICMP on your router, in order for me to set up monitoring as well?
 
I've enabled ICMP for echo replies.
I am currently not experiencing any latency issues at the moment, but I just thought of reaching out here to hopefully and finally find a solution.
Should I reply back here when it happens?
 
I've enabled ICMP for echo replies.
I am currently not experiencing any latency issues at the moment, but I just thought of reaching out here to hopefully and finally find a solution.
Should I reply back here when it happens?
Thanks , I will also monitor other clients in the area, to see if they experience the same issues as you are.

Yes please let me know when you encounter the issue again.

Regarding ICMP, unfortunately it still seems like your Public IP is not reachable, would you mind double checking?
 
Could you assist me with setting it up? This is the panel I am seeing

I have DHCPv6 enabled as well, since it somehow reduces my packet loss.
Is there potentially a conflict somewhere with ICMP?
 

Attachments

  • 1773765247384.png
    1773765247384.png
    68.6 KB · Views: 18
Could you assist me with setting it up? This is the panel I am seeing

I have DHCPv6 enabled as well, since it somehow reduces my packet loss.
Is there potentially a conflict somewhere with ICMP?
I believe you would need to change your source Zone to External, Destination zone would be WAN, and should be Echo Request,

Apparently there is also a setting on ubiquty , go to settings --> security , where you need to turn off "block WAN Ping"

Sorry I'm not familiar with Ubiquty, but this is what I found from some googling
 
This router is incredibly advanced and there is a ton of stuff here, but I have fixed the ICMP settings. Does it work now?
 
Read somewhere to make request and reply rules for pinging, working?
 
Read somewhere to make request and reply rules for pinging, working?
Still nothing unfortunately, I have already set up monitors to 3 other customers in your area tho, so just need to get yours to respond now. Not sure what were missing
 
Ahh wait, @Lepep have you tried rebooting the router after making the changes?, read here that it may be needed

But found this there as well, if you want to make sure its correct

Correct Firewall Rule to Allow ICMP (Ping):
  1. Log in to the UniFi console.
  2. Go to Settings > Internet > Firewall & Security.
  3. Create a New Rule:
  • Select the WAN IN or LAN IN section based on where you want to allow ping.
  • Rule Name: "Allow ICMP Ping"
  • Action: Accept
  • IPv4 Protocol: ICMP
  • Source: Set to Any (or specify an IP range).
  • Destination: Any or your gateway IP.
  • Enabled: Yes.
  1. Save the rule and move it to the top of the rule list.
  2. Apply Changes.
Optional Troubleshooting:
 
I see I have a newer interface to all these other sources, destination should be 'gateway' as there is no 'WAN IN'.
I have set new policies now, does it work?
 
Thanks, I will monitor and let you know what I find.
Please do let me know when you experience the issue again.
Once we have the monitoring data, will we escalate with TTC again to try and get to the bottom of the issue for you.
 
I am starting to see my latency climb, hovering around 20ms to 50ms, I expect it to potentially climb further past 8pm.

Edit: Calmed for now
 

Attachments

  • 1773770339053.png
    1773770339053.png
    14.5 KB · Views: 11
I am starting to see my latency climb, hovering around 20ms to 50ms, I expect it to potentially climb further past 8pm.
Can confirm,

So I also added 4 targets from another FNO on the same Access Router as you, just to rule out our equipment as well
Seems only the TT Connect targets in your area are seeing the latency increases.

From what I can see on your ticket history, TTC mainly focused on making changes on your line, where they should have focused more on the exchange/backhaul side of things.


Will escalate with them in the AM once we have more data1773770538337.png
 
Top
Sign up to the MyBroadband newsletter
X