The plot thickens!

 

This is London > Edinburgh over LINX LON1 and private peering in
Edinburgh. Interesting how it just seems to be afternoons, and
not weekends.
This pattern continues for a few weeks beforehand. 

- A

On Mon, Oct 15, 2018, at 8:51 AM, Chris Malton wrote:
> Hi Alfie,
> 
>  It appears that 70-80ms to their router is what I was seeing from my
>  office in Southampton, not through it from Gatwick (I have too many
>  damned graphs now!).> 
>  This was last night:
> 
>  
> 
>  You can safely ignore the packet loss around 2150-2205 - that was
>  a.n.other incident entirely (one of the ones I'm diagnosing for said
>  customer)!> 
>  Not seeing anything today yet, however here's the above destination
>  for the last 10 days:> 
>  
> 
>  Do these match with anything you're seeing?
> 
>  Regards,
> 
>  Chris
> 
> 
>  On 14/10/18 23:10, Alfie Pates wrote:
>> I’ve been seeing some unusual latencies recently too; can you share
>> some graphs and I’ll compare with mine!>> 
>> - A
>> 
>> 
>> On Sun, Oct 14, 2018, at 9:22 PM, Chris Malton wrote:
>> > Hi all,
>> >
>> > I've been diagnosing a load of issues with a client service
>> > recently,>> > and part of that's led to adding Smokeping for literally
>> > everything.>> >
>> > Alas, that's turned up rather more than I had expected it to -
>> > some of>> > which doesn't appear to be in my control or my DC's control or 
>> > my
>> > client's control either - and it's skewing some of the
>> > measurements!>> >
>> > Of an evening, and at weekends (i.e. peak time internet usage)
>> > for at>> > least the last week or so, I've been seeing very high latency on
>> > round>> > trips from my rack in 4D in Gatwick to destinations in
>> > JISC/JaNET.  That>> > seems to be the only affected route for me, but is 
>> > the only one
>> > via TI>> > Sparkle that I've currently found.  There's a little bit of
>> > packet loss>> > in the mix as well (~0.1%), but nothing I'm immediately 
>> > concerned
>> > about.>> >
>> > I've dug more into this over this weekend, and have identified that>> > 
>> > AS6762 appears to be where the issues lie - specifically, the
>> > moment>> > traffic hits TI Sparkle's LINX router (2001:7f8:4::1a6a:1) I 
>> > see an>> > 80-90ms latency increase at peak times.
>> >
>> > Is anyone else seeing high latency on routes via TI Sparkle
>> > across Linx>> > and/or is there a friendly contact at TI Sparkle I can 
>> > discuss this>> > with?  (Separately, I've asked 4D to peer directly with 
>> > JISC on
>> > LINX ->> > but that might take some time to get sorted)
>> >
>> > Thanks in advance.
>> >
>> > Regards,
>> >
>> > Chris Malton
>> >
>> >
>> 
>> This email has been scanned by the MailScanner e-mail security
>> system.> 


>  
>  

Reply via email to