6 Jul 2013
I have recently started using Windows Azure for a few of my websites, and they recently announced Alerting and Autoscaling as a preview feature.
I have used Pingdom for my checking of up-time, and still will do, but I thought I would try the new alerts for uptime to see how they compare.
First you need to log in to Azure Management. The website/service you will be monitoring will need at least one endpoint setup under its settings:
I have setup the homepage URL and three locations worldwide. These then show under the monitor section with response times, which is great for seeing any slow downs over a period and what kind of response time people on the otherside of the World are getting.
Now for alerts, once you have an endpoint setup, go to the Settings section at the bottom:
Then click Add Rule at the bottom:
Now type a name, and optionally a description. Choose the subscription, service and service name. I chose website and makit for this website.
Click the next arrow and you will be given the choice of which metric. Choose an Uptime metric from one of the monitor endpoints you have setup. I chose one close to the UK because that is where I am based.
You then need to set the condition to less than (<) and add a value. I want my website to be 100% available and any downtime is bad, which is how it should be. Therefore I set the Threshold Value to be 100. The Unit is fixed as a percentage sign.
I left the alert evaluation window to be the average over the last 15 minutes because I'd like to know within 15 minutes of the site going down, and not 60 minutes.
Finally I chose to email the alert by choosing "Send an email to the service administrator and co-administrators".
Finally, make sure the Enable Rule is checked and click the Tick.
Any alerts will be emailed with a submit of "ALERT ACTIVATED" from a mail.windowsazure.com email address. You can also go to the alerts section in the management console and see the graph of uptime over a period.