I have an MVC based web app running on Azure. The CPU performance of it has been very predictable over the past five months. However, over the past 24 hours, and most recently, from 1:00 pm to 1:30 pm Eastern time, today, in the USA, I have had CPU spikes nearing 100%. The image below, which is for the past 7 days shows this.
This CPU spike is not coming from my app or my users. There has not been an abnormal increase in users, user activity or queries. I also checked Google Analytics to see if perhaps my site was getting hammered by random users etc. It showed nothing out of the ordinary.
There also was a corresponding huge jump in data going out of my site, which is highly unusual. The second image shows data egress for the past week. However, as I said, I checked my Azure SQL Database Query Store and it shows absolutely nothing out of the ordinary. Furthermore, my DTU percentage never even neared 100% during this time, which it certainly would have if this much data was pulled from the database.
I have basically ruled out anything amiss on my end. Is there some way I can check to see if there were issues with Azure causing this?
If you are suspecting an underlying Azure platform issue, both Azure Service Health and Azure Resource Health are useful resources to determine if you are being impacted by platform issue.
Azure Service Health provides personalized service health information when Azure platform issues impact your resources.
https://learn.microsoft.com/en-us/azure/service-health/service-health-overview
Azure Resource Health provides visibility into whether your Azure resources are healthy or unhealthy.
https://learn.microsoft.com/en-us/azure/service-health/resource-health-overview
For a list of supported Azure resources, you can refer to this article which also describes the set of health checks being performed.
https://learn.microsoft.com/en-us/azure/service-health/resource-health-checks-resource-types
Related
There has been a previous question on this and the accepted answer was Azure Elastic Job agent. The problem I have is that the feature is in preview and it still lacks a lot of functionality like diagnostics and alerting. I also find it to be very unreliable as job get randomly cancelled because of service restarts.
Azure Automation Accounts also works, but it only has a execution/running time of 3 hours. So if your maintenance takes more than 3 hours, this is not an option.
I have previously developed my own application for doing this, but the maintenance and management of this can become a headache.
Another alternative could be to just leverage Azure Data Factory perhaps, but this is a route I have not yet followed.
So what are people actually using to do long running maintenance against Azure SQL Databases that has enough diagnostic information in case something goes wrong and has at least some level of alerting?
PS: The database I need to do maintenance on is not small.
I have an app service plan which is hosting 2 Web APIs, the issue I am facing is that I am unable to view details such as: CPU Usage, Memory Percentage, Requests, Average Response time etc.
These can be found under the Overview tab for both App Service and App Service Plan but no data is being recorded, even if I retrieve data for the whole week rather than the last hour only.
I have also confirmed that I am hitting the correct App hosted on the correct Plan. Have I missed anything? Do I need to enable something?
I have also generated around 20k requests in the last few hours so I expect something to show up.
I have an Azure Cloud Service published at Microsoft and it's draining all my credit!
Payment
Pay as you go
Service resource
Minimal resource, 1 SMALL web role and 1 SMALL worker role.
I knew Azure wasn't cheap, but this is just too much. Currently my monthly cost is just under 80 USD. The only person that use this service is me, noone else, and I barely use it. So the cost is just for the upkeep.
Is this normal?
70 bucks a month!?
How much does it cost for YOU?
What Microsoft support told me
I am afraid the Cloud Services has a fixed price, and I am not aware
how it could be lowered. Maybe you want to check on how the service
itself could be tweaked to get it working as per your needs. You may
want to go through the Community Forum for that.
Community = Stackoverflow, so here I am!
If I look at my Azure subscription page I can see that it's the:
CALCULATING HOURS - Europe, Western
That is taking all my hard earned money. My service also uses SQL, storage and cache but, if I understand it correctly, these are not the cause for my expensive bill.
Before I leave you to it I just want to say that I can't use a simple web app because of my requirements. I know web apps are super cheap, but in this case I must use a cloud service..
Thank you
Update
I found out I was using A1 (small) and not A0 (extra small). The instance type for a cloudservice can be set in the servicedefinition file.
It's sad that not even Microsoft themselves could inform me about this.
Web and worker roles are like dedicated VMs if they are on, they will cost you money.
You can do one of two things
1) Stop the machines when no one is using them ( say in the off business hours). I am not sure if this is possible to do or not in your case. But if it is possible, you can run a small script to start/ stop the roles. You can even do so via apps on your phone. For example - https://itunes.apple.com/us/app/azure-management/id826446897?mt=8
2) Move to Azure Web Apps and Azure Web Jobs - Both these services are "multi-tenant" and cost much less and in fact offer a free tier. If and when you need to scale, you can always scale as your need
Hope this helps
Hi I have a web app deployed as a Cloud Service on Windows Azure. Now I am performing some load/stress test against this app. In the Azure Management Portal I have configured the web role to scale automatically when the CPU goes over 40%.
I start the tests with only one instance of this web role. As the test progresses, I have set the number of concurrent users to increase over time up to 2000 users.
After I start the test, I connect via remote desktop to the web role instance on Azure and I monitor the CPU usage. After 10 mins or so, the CPU is constantly at 100% (and in fact my requests in the test take a very long time to complete) but if I check the CPU of the very same web role on the Azure management portal it says 1, 2 or 6, there was a peak of 70% but it sunk back immediately (but never the values I see in its task manager when I am connected in remote desktop) or even does not display any value (I go to the dashboard page of my cloud service), which means the graph is not updated any more.
Furthermore, and this is the point of my question, NO SCALING of the web role instances is performed whatsoever.
Any ideas where/what I am missing? Feel free to ask if my explanation is incomplete.
Autoscaling on the CPU metric for a Cloud Service or Virtual Machine doesn't occur as fast as you are expecting (~10+ minutes). In this scenario, the CPU metric is averaged across all instances of the services for a period of 1 hour. Therefore, your autoscaling actions will not be immediate.
You can read more about this and some recommendations for configuring your autoscale settings here.
If you want to tighten this up a little more then take a look at this post where I show how to set the TimeWindow using the Monitoring Service Management Library. You may be able to get closer to what you want taking this approach.
A few things to consider:
1) As Rick pointed out, by default CPU is taken at an hour average
2) If you start at only 1 server, and then autoscale up to 2, your first server will get yanked out of load balancer during the scale operation. You should really always have a minimum of 2 servers at all time.
3) Feel free to check out AzureWatch (link in my profile).. it was designed to perform decently advanced scaling scenarios and allows you to configure scaling rules without touching APIs
My 3 months trial account Windows Azure has already been disabled in 3 weeks. How's that possible? I had nothing on my project, just a simple Asp.Net web page.
I don't think anybody knew my page and made constant requests.
I can't find a statistics section on Management Portal to check what was my traffic...etc.
Does anybody know where I can check my Hosted Service's statistics?
Thanks
The trial provides 750 compute hours monthly. Once you deploy your app, the meter is ticking. That is, as long as something is deployed, it's a metered resource. Whether consuming 0% or 100% cpu/network/memory, you pay hourly.
Now: If you deployed a single Small instance for your asp.net site, you shouldn't have consumed 750 hours in 3 weeks. Is it possible you deployed with Medium instances? Or deployed with 2 Small instances? Do you have more than one Role in your deployment (since each would have at least one instance)?
One bit of advice I give, when doing dev work: at the end of the day, when you're not actively working on a project, delete the deployment (you can always re-deploy to the same place later). This helps save tremendously on consumed hours.
You aren't charged based on access to a hosted service (i.e. compute instance). As long as you have something deployed there, you are charged. No one can ever go to it and it still costs you money.
At the portal, you have to delete whatever hosted services you have there in order to preserve credits. For example if you had up both a Production and a Staging instance, they would charge you compute hours for both. You have to actually delete the instances in order to conserve compute hours on your bill.
As for stats, the only way you can get access stats that I know of is by using the azure diagnostics features. They used to have a lot more detail on their bills (in / out transfers, etc) but the bills are a lot shorter now.