I never got that ASN. I still have the SLA paperwork sitting in my drafts folder. Curious if the bimodal latency pattern uma saw ever got explained properly by the carrier or if it was just another "resolved itself" closure.
Never got a reply from anyone on that. Thread died and the issue cleared itself after about six hours.
I still have the MTR logs somewhere if anyone cares. The bimodal latency split at 15 ms vs 85 ms was the interesting part.
Did this ever get resolved properly. Still waiting on which carrier was at hop 4 so I can check if SLA remedies were actually paid out. 246 days is a long time for a transit issue to go undocumented.
That bimodal pattern you saw, was it asymmetric or did both modes shift together. Asking because I've seen similar on KnownHost peering and want to know if this was the same root cause or just same symptom.
I never got that ASN. I checked my notes from that week and the latency pattern cleared around 14:00 UTC same day, but I never saw a root cause published. Would still like to know if Contabo or OVHcloud ever updated their SLAs after this.
Member OP
Harold, sorry lah this thread so old already. I check my old MTR log, hop 4 was 129.250.x.x that time. I think Level3/CenturyLink but not 100% sure.
I never did get an answer on this. I ended up moving that OVHcloud workload to HostHatch's AMS3 region and the problem vanished overnight. Still curious if there was ever a root cause published.
Thread went cold but I'm chasing a similar pattern on a KnownHost circuit into DE-CIX this week and wondering if it's the same upstream.
I never did get an answer on this.
Member OP
Harold, sorry lah this reply so late. 440 days already, I totally forget about this thread.
I never got the ASN properly because the MTR cleared up same day around 14:00. Loss just stopped, no explain also. I think was Level3 side but cannot confirm 100%.
Did you ever find out which carrier got the SLA hit? I still with Contabo FRA1 but their status page same like OVHcloud, always green even when got issue. Very useless.
443 days and I still have the MTR logs sitting in a tarball somewhere, probably next to my OVHcloud refund request.
