Appharbor performance porblems persisting
In a former post I pointed out that I would have frequent
timeouts trying to reach my site on Apphabor.com. It was suggested
it could be due to some general network bottleneck. This morning
(11:00 local - GMT-3) I once more did a tracert for appharbor.com
and heroku.com (which are supposedly both on ec2 servers). And as
always, although neither heroku.com is like lightning fast (from my
location: Belém - North of Brazil), it still performs much
better than appharbor.com.
I think the top 2 requirements of a hosting service (free or paid)
are availability and speed. My question: are there plans to improve
on these requirements on a global level, or will appahrbor continue
to focus mainly on US users, as seems to be the case currently.
As it is, the service has no real practical use for people in my
situation, for now.
Here go the 2 tracerts:
C:\Documents and Settings\Administrador>tracert
appharbor.com
Rastreando a rota para appharbor.com [184.72.232.223]
com no máximo 30 saltos:
1 2 ms 1 ms 1 ms 10.1.1.1
2 142 ms 82 ms 44 ms 200217089016.user.veloxzone.com.br [200.217.89.1
6]
3 92 ms 63 ms 63 ms gigabitethernet10-2.81-cto-pa-rotd-03.telemar.ne
t.br [200.217.30.94]
4 104 ms 75 ms 111 ms 200.223.44.193
5 206 ms 170 ms 327 ms 200.223.50.62
6 197 ms 275 ms 203 ms ve433.ar9.NYC1.gblx.net [64.215.184.21]
7 358 ms 481 ms 347 ms po1-20G.ar4.DCA3.gblx.net [67.16.132.141]
8 524 ms * * po1-20G.ar4.DCA3.gblx.net [67.16.132.141]
9 * * * Esgotado o tempo limite do pedido.
(...)
30 * * * Esgotado o tempo limite do pedido.
Rastreamento concluído.
C:\Documents and Settings\Administrador>tracert heroku.com
Rastreando a rota para heroku.com [50.16.233.102]
com no máximo 30 saltos:
1 1 ms 1 ms 1 ms 10.1.1.1
2 208 ms 126 ms 235 ms 200217089016.user.veloxzone.com.br [200.217.89.1
6]
3 236 ms 119 ms 237 ms gigabitethernet10-2.81-cto-pa-rotd-03.telemar.ne
t.br [200.217.30.94]
4 * 92 ms 123 ms pos10-0-ald-ce-rotd-02.telemar.net.br [200.223.2
54.21]
5 161 ms 467 ms 290 ms 200223045158.host.telemar.net.br [200.223.45.158
]
6 244 ms 431 ms 449 ms ve433.ar9.NYC1.gblx.net [64.215.184.21]
7 299 ms 268 ms 309 ms po3-20G.ar4.DCA3.gblx.net [67.17.107.198]
8 235 ms 220 ms * po1-20G.ar4.DCA3.gblx.net [67.16.132.141]
9 * * * Esgotado o tempo limite do pedido.
(...)
15 235 ms * * ec2-50-16-233-102.compute-1.amazonaws.com [50.16
.233.102]
16 275 ms 204 ms * ec2-50-16-233-102.compute-1.amazonaws.com [50.16
.233.102]
17 243 ms 174 ms 202 ms ec2-50-16-233-102.compute-1.amazonaws.com [50.16
.233.102]
Rastreamento concluído.
Regards,
Jaime
Discussions are closed to public comments.
If you need help with AppHarbor please
start a new discussion.
Keyboard shortcuts
Generic
| ? | Show this help |
|---|---|
| ESC | Blurs the current field |
Comment Form
| r | Focus the comment reply box |
|---|---|
| ^ + ↩ | Submit the comment |
You can use Command ⌘ instead of Control ^ on Mac
Support Staff 1 Posted by friism on 18 Mar, 2011 06:44 PM
The only comparable numbers in the traces above are for the gblx.net host in New York and the hosts closer to you (this is because the AppHarbor servers are not configured to respond to ICMP requests) and they can not be used to compare AppHarbor and Heroku network performance. Here's short guide to understanding tracert output.
We are commited to providing fast and reliable service, but we are also constrained by the infrastructure available to us. We expect to roll AppHarbor out to additional locations as we see customer demand spread.
friism closed this discussion on 18 Mar, 2011 06:44 PM.