I am new to web site development, and I have been facing some troubles to find an answer to this question. I use Heroku, nonetheless, the domain is not completely personalized, and it sends the dynos to sleep if no one accesses the page, therefore, the first user will always face delays o page reload. Further, it does seem to appear on Google. Can I use Heroku to create a functional web page? if not, which option would you suggest? I want something simple and efficient: i.e., I want to concentrate on coding my site, not coding details to make it run. Detail: I am using Node Js (MEAN stack).
A verified Heroku account gets 1000 dyno hours every month. A month has 720 hours. If you are deploying only one web app you can permanently host a website there.
There are free solutions that ping your web page in regular intervals to prevent it from sleeping. E.g. you could use the New Relic add on.
....and the site doesn't respond in that time at all. Some responses even fail.
This is giving a horrible user experience where the site suddenly stops reacting for the people that are online at that moment.
The site is completely warmed up and responsive through the temporarily url.
Here is the log:
Is there anything to speed this up?
Is this normal behaviour or is this an error that should be reported?
From the log you posted it seems that the operation that consumes a long time is "ChangeDeploymentConfigurationBySlot".
An azure configuration change by default triggers the restart of the role instance (which might explain why your clients are experiencing downtime).
You can get further details in this blog: https://alexandrebrisebois.wordpress.com/2013/09/29/handling-cloud-service-role-configuration-changes-in-windows-azure/
I have a number of small MVC apps deployed as Microsoft Windows Azure websites. This has been working for several months.
Yesterday I rolled out a new one, and the deployment was unremarkable, everything worked fine. But a couple of hours later, access to the site was unavailable. The symptoms were that when the browser tried to navigate to the URL for that site, it would try to load for several minutes and then just give up with a completely blank page.
I attempted to stop and restart the site, and it worked once, but the symptoms came back several minutes later. Then I tried to stop and restart, and it didn't work.
I deployed the identical app to three additional URLs. Again, immediately on deployment, they all work fine, however, they fail at some interval in the future. They seem to not all fail at once. Sometimes restarting the site will fix the problem, and sometimes not.
IMPORTANT: If I wait for some period of time, the site may start to work again on its own.
However, deploying four versions of the app so that our users can go to a backup one if the primary one is not working is not optimal.
Any words of wisdom as to how I might go about debugging this?
ADDITIONAL INFO NOV 25, 2013:
When sites are failing, the IIS logs show either 500 or 502 Internal Service Errors. Our own MVC code is never hit, not even app_start.
You can start by checking the logs and remote debugging
http://www.drdobbs.com/windows/azure-sdk-22-supports-visual-studio-2013/240163499
Are the apps working locally?
Might not be the same problem, but from time to time our Azure instances will get the blue question mark of death as a status.
The reason we found out was that Microsoft will do upgrades on instances from time to time. If you have just one instance in a cloud service/role, then from time to time they will do maintenance and during that time it will be dead.
I have confirmed this with their support.
The only way to get around this that I know of is to create two instances. Then Microsoft guarantees ~99% availability.
Of course I also confirmed with them that this means twice the cost. =/
If that's not the issue I would enable RDP and get onto the machine to see what the problem is. Microsoft has these tools to help debug problems: http://blogs.msdn.com/b/kwill/archive/2013/08/26/azuretools-the-diagnostic-utility-used-by-the-windows-azure-developer-support-team.aspx
First, you should always run multiple instances of your web role with more than 1 upgrade domain. This is configurable in the service definition (CSDEF). Without this, you don't get an SLA from Microsoft, so you can't really complain that the VMs go down.
Second, to figure out what might be going on with these boxes, you should have both logs (my preference is to roll my own with page blobs or table storage), AND you should always have RDP access to a pre-production environment (production as well if you're not too fussed about security). Once on the box, look through the event viewer for errors.
Third, when an outage occurs check out the azure service dashboard (http://www.windowsazure.com/en-us/support/service-dashboard/) for outages.
Lastly, contact Microsoft support. It may take a few hours, but they are pretty good.
That it is happening repeatedly and for extended periods of time (more than 5 minutes), I would be there's something wrong with your hosted service. Again, RDP in and poke around. Good luck.
To debug your sites try to enable diagnostic logs:
http://www.windowsazure.com/en-us/develop/net/common-tasks/diagnostics-logging-and-instrumentation/
Another nice way to look around your site is using the debug console:
https://github.com/projectkudu/kudu/wiki/Kudu-console
Older version of site is appearing on some computers but at the same time I can able to see the new version of site.
I have done with the clearing of cookies and caches. But problem still occurs.
Any suggestions please.
You have to wait untill all DNS servers will update their domain records - it's called DNS propagation time.
Some computers use different DNS servers (internet provider's, Google's, the nearest one), and it takes some time to update them all, as they are csheduled to update in different time.
Usually it's done within 24-48 hours. It may take longer sometimes, but usually it's done in no more than 6.
I have a website that seems to have a really slow dns look up. It can take anywhere from 2.7 sec to 18 sec. But I was playing around with it and went to the page and it took 48.91 sec to do the dns lookup and all of the time it was "waiting" Can someone explain what the different stages are and if it's my website that is causing the waiting problem or if its the hosting company? The website is a Joomla website and the hosting provider is NetworkSolutions.
Here's an excellent article and talk by Tom Daly regarding speeding up DNS lookups and decreasing TTB (Time to Byte).
http://dyn.com/blog/time-to-first-dns-query-response-performance-behind-the-byte/
http://youtu.be/iCfuKZUyTko