John Fan's dashboard

John Fan replied on January 21, 2009 00:21 to the question "Is DevjaVu still being actively maintained?" in DevjaVu:

Thanks... what about the plans for continuing the maintenance of this service for the next few years?

John Fan asked a question in DevjaVu on January 20, 2009 23:06:

Is DevjaVu still being actively maintained?

Hi, I am thinking of using DevjaVu... is the site still working? Can you provide guarantees on availability?

John Fan replied on December 22, 2008 21:22 to the question "Can you report # of mongrels on the Weekly Report?" in New Relic:

for rightscale, this number is NOT fixed, and can vary from day to day

also, in live deployments, certain hosts or mongrels can go down, so this is a way of observing that as well.

but I think max (e.g. per day) is sufficent for the weekly report.

John Fan asked a question in New Relic on December 22, 2008 09:04:

Can you report # of mongrels on the Weekly Report?

on the Weekly Application Performance Report, is it possible to add a column for the number of mongrels?

John Fan asked a question in New Relic on December 15, 2008 08:16:

Can you provide the "referring URL" for error traces?

Is it possible to see the referring URL for an error trace in the case of URL not found? In particular, the problem is that we are getting a number of banner ads that point to the wrong URL. We see it as a "missing action / URL not found" error on our server, but we have not been able to identify which ad network is responsible for this error...

John Fan replied on June 19, 2008 16:29 to the question "server is no longer reporting to newrelic.com" in New Relic:

Thanks for addressing this issue. I'll disable it on the staging machine and try it out on production again.

John

John Fan marked one of Lew's replies in New Relic as useful. Lew replied to the question "server is no longer reporting to newrelic.com".

John Fan replied on June 19, 2008 07:43 to the question "server is no longer reporting to newrelic.com" in New Relic:

Lew,

Sorry, but I will have to give you an earful:

As soon as the limits started being enforced (about 5 hours ago), my app started to slow down. Even after I upgraded to the paid version (and 8 mongrel-allocations were placed on the production server, which actually has 10 mongrels on it), the performance was terrible. Finally, we decided to turn off New Relic (enabled: false), and the app is performing perfectly.

In short, when New Relic enforced the limits, it not only killed off the monitoring but killed my app as well. In addition, even with the paid version, if not every mongrel is being monitoring, the app is also unusable.

I'm not sure I can continue using this service if it risks bringing down my app.

In any case, can you please move all 10 mongrel-allocations over to the production server (ec2-75-101-232-94.compute-1.amazonaws.com)?

John

John Fan replied on June 19, 2008 06:38 to the question "server is no longer reporting to newrelic.com" in New Relic:

That's too little!

There should be 10 (not 8) from the main server. There should be 8 from the "staging" server are only used to test the deployment and aren't important, but it would be good to have the full access to newrelic stats to make sure it is working before I deploy to production.

John Fan replied on June 19, 2008 02:58 to the question "server is no longer reporting to newrelic.com" in New Relic:

Hi Lew,

Ok, I just signed up for a paid account. Can you help me resolve this issue?

Thanks,

John

John Fan replied on June 19, 2008 02:27 to the question "server is no longer reporting to newrelic.com" in New Relic:

Hi Lew,

Yes this is a free account.

When you say 2 mongrels, do you mean two app servers or two mongrel instances?

We are currently using 10 mongrels on 1 app server, and it was working up to a few hours ago.

John

John Fan asked a question in New Relic on June 18, 2008 23:04:

server is no longer reporting to newrelic.com

Hi, New Relic was working fine for me, but it has stopped working (stats are no longer visible on newrelic.com)! What's strange is that is still working for the staging server, which should have the identical configuration...