bro how is your unifi today?

me melaka area also quite bad, everything is very slow.

Even trace route show that the second hop ping is around 100MS from customer CPE to TM router. Tons of packet loss as well and speedtest to tm server only get 20M out of 300M package.

Lag a bit on lazada, qq already 72 hours without disconnect, fine on cloudflare SG cdn, bilibili smooth on live.

ur gateway is different from me, hmm...
But yes, going to SG always reach 10.55.100.246, so your gateways is busted...

Me last time is 10.233.97.55 -> dropped, but yours is router -> 175.137.x.x high latency, I think it can be your area overloaded over the same fiber signal, or your fiber getting bended?

Check for fiber signal if you know how to do it or easy to done, can try disconnect and avoid 175.137.x.x IP address.

fiber optic cable no bended and the signal is ok, not sure why it can happened suddenly.

yea tried to switch to different IP block as well, but all same. So already asked tm open report to investigate

fiber optic cable no bended and the signal is ok, not sure why it can happened suddenly.

yea tried to switch to different IP block as well, but all same. So already asked tm open report to investigate

Try ping and traceroute this cloudflare ipv6 dns : 2606:4700:4700::1111

Lag a bit on lazada, qq already 72 hours without disconnect, fine on cloudflare SG cdn, bilibili smooth on live.

ur gateway is different from me, hmm...
But yes, going to SG always reach 10.55.100.246, so your gateways is busted...

Me last time is 10.233.97.55 -> dropped, but yours is router -> 175.137.x.x high latency, I think it can be your area overloaded over the same fiber signal, or your fiber getting bended?

Check for fiber signal if you know how to do it or easy to done, can try disconnect and avoid 175.137.x.x IP address.

Hi what is the software on the ss u provided? My tracert on Windows is not working as usual, seems like TM block it or some sort.

Hi what is the software on the ss u provided? My tracert on Windows is not working as usual, seems like TM block it or some sort.

IP pool rarely makes a difference now. I’m in JB as well (Mt. Austin), and have been assigned all the IP ranges you listed, including the obscure ones like 124.13.xxx or 1.32.xxx and it seldom makes much of a difference. On bad days, basically there will be a period every hour that lasts anywhere between 5 - 30 minutes where certain peering routes get overloaded and TM will do traffic shaping or load balancing of some sort (I don’t know the exact technical term, sorry). The only difference your assigned IP pool makes is the time it occurs, but it almost definitely will occur once every hour for all these IPs.

Like you say, a VPN helps, but there are limitations. If the route to your VPN is through one of these overloaded peering providers, then it makes no difference. For example, for your VPN’s SG servers, it would depend on which datacenter they’re housed in and what peering hops are taken back to your TM IP address. In my experience, M247 and DataPacket/DataCamp servers in certain datacenters rely on PCCW, which is completely f@#ked when this issues happen. M247-Level3 is much more reliable. If what you say is true and OVPN relies on OVH, then you’re already in a quite good situation since (according to a poster here) OVH has a private peering agreement set up with TM.

Anyway, it’s a good idea to do proper traceroutes to any commercial VPN provider’s servers in the region before you commit to a subscription. Even then, it’s better to pay a bit more for a monthly sub rather than discounted yearly deals, since TMs peering situation may change at any time. Obviously rolling your own on a VPS is also great,  ut those have their own limitations like reduced privacy (because most require CC details) and data caps.

Level3 (left) vs PCCW (right) to SG M247 servers during the usual koyak period just now.



Typical PCCW

Hmm, another crash at 0 AM and now heavy packet drop on 10.233 to 10.55, most oversea line force disconnect, guess another CGNAT issues?

Strange that it happen sharp 0 AM again lol, guess someone setup a bot or spider to send a lot of traffic + TCP half-open line during 0 AM.

Not sure if related but I have restarted my router 3 times in the span of about 10 minutes to get it to change gateway and get a new IP, my twitch connection has been very bad, mainly connection not being consistent, dropping from 1080p60 randomly to 144p. Same with my Apex games, packet drops and high ping (Not sure if this is my side or EA's side). But it's been really bad for my connection in general. Feeling like CGNAT gateway recently has been getting congested after 12AM straight.. 

And another is, the Malaysia going out/in is crashing.

Like lazada now, the static CDN server is at Malaysia Akamai, it take 21s for loading, and telling the browser it is CDN server loading/downloading from main server at SG.
So, even local CDN making it worst....  

Wow that's interesting.. TM should really figure out what's going on with their CGNAT implementation cause the connection issues has been getting worst lately it looks like 

very headache this unifi...
cost so expensive yet many problems...

Yes, even lazada on MY akamai affected, which using TM IPv6 as its CDN MY server...
So this CGNAT issues are forcing people sub to public IP plans, which is 300M if I am correct.

Else you gonna seek for old CCTV that need public IPv4 for access.
---
Anyways, now the network is a mess for me, since the break down around 0:15 AM, 95% connection got packet loss issues even just to Lazada.

Lucky for lowyat which using CloudFlare is HTTP2/HTTP3, so far those reuse single connection with multiplexing don't affected too much, but those need recreate new connections for most new request are failing like crazy.

Worst as lazada CDN to Lazada SG backend server also got issues, double the issues, triple the trouble…