I have a Flask Application (python for the backend + js/vue for the front) that is running on an Azure App Service. I face very bad performance : lot of time during the day, the application takes minutes to responds. I cannot understand why, is this App Service is very bad ?
Regarding the App Service Plan, I have the P2V3 (220€/month, 16go and 4vCPU), I have also a SQL Database to store information.
I don't know if I can provide more information, but it is really impossible to use it on a long term, for example here some of metrics of the "slow down" issue :
So it is related to Azure issue (and for example can be solve moving on AWS), or to my application ?
If there is any information that I can provide to help identify the issue I can provide, I am not an infrastructure expert so I am not really able to identify myself what it is important to check.
Thanks
#kilag, thank you for confirming this. Looking at the graph you provided, this seems to happen intermittently, correct? Has this behavior been occurring since the start or did it recently start happening after a recent deployment?
Are there any other external dependencies beyond the SQL DB? Such as a CDN?
I would suggest following #azMantas's advice to try out the diagnose and solve blade. This is the same tool that support engineers use when you purchase a paid technical support ticket. More information on that here. This is the quickest way to uncover a potential issue.
Please let us know if there are any insights provided such as a storage volume failover that might be reported in the diagnose and solve blade.
If you see nothing in the diagnose and solve blade, please reach out to us at azcommunity#microsoft.com with the subject ATTN: Bryan and in the body of the email, include your azure subscription ID so we can assist you further. We look forward to your reply.
Related
I am experimenting with the Azure Computer Vision service and it is working well.
The only issue is that I want to be able to process 6,000 files relatively quick.
Now Azure imposes a 10-Transaction Per Second limit which is fine, but how can I scale the service?
I was thinking of spinning up multiple instances, but it seems like an ugly way to solve this issue.
Please let me know the best way to tackle this issue.
Thanks!
Please contact Azure support for scaling your Azure Computer Vision resource.
You can directly ask the support to have an upper limitation.
Submit a support request from Azure for that:
I have an Azure WebApp that continually reports "Your app experienced failure(s) due to a transient storage access issue." The suggested solution is "Explore Local Cache feature for your web app." but my webapp exceeds the maximum storage (3GB) for this option.
The problem mostly occurs between midnight and 6am in the morning when the site is LEAST active, but there seems to be an increasing number of occurrences during the day.
What are the underlying causes of this problem? Is this something to do with my WebApp or is it the Azure Infrastructure. In either case, how do I determine the underlying issue(s) and resolve?
"Your app experienced failure(s) due to a transient storage access issue."
The Web Apps environment provides diagnostic functionality for logging information from both the web server and the web application. You could try to enable Logging and check the logs that generated within that period of time.
According to the error, it seems that a temporary issue causes app failure, and it suggests enabling local cache. You could follow the suggested solution and make sure if it helps resolve issue.
Besides, you could try to scale your web app (which would take additional charge) and check if it could mitigate the issue.
Updates:
As we know App Service offers a shared, persistent storage for the application. Maybe something wrong with the shared storage when the instances in farm access the storage, which may be the cause of the issue.
To determine the underlying issue, you may try to enable diagnostics logging for your web app. This should provide more information on what is happening at the storage level and what kind of activity is going on.
We use the Northern Europe Azure Data Centre. We use SQL Azure Web Edition with Websites. Today we seem to be getting quite a few SQL Azure disconnections with error code 10054:
Symptom:
A transport-level error has occurred when receiving results from the server. An established connection was aborted by the software in your host machine
Mitigation:
Implement retry logic in your application
What is frustrating about this is that I cannot find any MS Azure Alerts to an issue. I suspect their SQL Azure systems will be ringing alarms. I have looked in Azure Status and Management Services.
Are there other places I can looks for these alerts, and also what is the best way to notify MS that there is an issue.
Finally how are folks monitoring SQL Azure for issues, throttling, disconnects etc? I have come across cotega.com.
P.s as a side note, I do understand the importance of putting in "retry logic", and but cannot happen quickly, and my code could be improved in this respect. However something does seem wrong with SQL Azure today. Up until now, the service has been by and large great.
EDIT 1:
Error Code:
Error Number:10054,State:0,Class:20
EDIT 2:
I think the moral of this story is to not underestimate the impact of throttling and scaling out, rather than scaling up, of SQL Azure services, and the importance of implementing retry logic/ MS Transient pattern logic. Still, this variable performance is a little frustrating, but perhaps this is the cloud for you, although with the new Data Tiers you can pay for more predictability.
EDIT 3: Interestingly SQL Azure is not performing more efficiently with no errors being noticed. So an issue with Azure. However difficult to identify apart from application error logs. I use Elmah.
Reconfigurations happen when upgrades are rolled-out to the service or when databases are load-balanced to avoid noisy neighbor issues.
The view sys.event_log in the master DB has information on why the disconnects happened (reconfiguration vs. throttling, etc.).
Service outages are reported in the portal.
I am a newbie to web development
I would like to host my site in azure.
There are so many subscriptions plans.
So which subscription is reasonably good and give me price details of that?
Thanks in advance
Windows Azure has few types of hosting. For a website you might want to look at the following -
Web Sites - You can host right away without modification of your existing project.
Cloud Services - I used this, but it requires changes such as Caching.
Here is the calculator based on your need.
FYI: Rule of thumb is you need a least two instances in Production to minimize the downtime.
If you are a newbie , I would strongly suggest using azure websites for now, and you can always move to a custom solution using webroles/caching Etc later if you feel it doesn't cater all your needs..
Azure websites pricing can be obtained from here :
https://azure.microsoft.com/en-us/pricing/details/web-sites/
Again on what parameters would you choose the right package, you are the best judge for that since you know what traffic are you expecting and how much memory etc you need
I have got two database (SQL azure) in North America, I'm getting error that the applications can't Access to the server.
I didn't update the application or database, so I supose that there are a problem with sql azure service. How can I notify Microsoft of this problem?
To 'notify' Microsoft, you have very limited options.
They have their own support forums.
They have their own support ticketing system, that costs a pretty penny but is the fastest way to get their attention.
They have their dashboard, which in my own experience is terrible. It is not a true representation, ever. The updates are very very late.
You also have StackOverflow - but there will be little that we can advise on if there is a problem on the Azure infrastructure side of life.
To help aid in your Azure support woes, I would suggest you get an account with Pingdom and get MetricHubs for your subscription. These will help in showing what goes down, when, how often, and for how long. It can help show if the problem really is in your application or not.
I would also ensure you have diagnostics set up, and log everything you can.
Many many people forget or don't know about the transient error problems. Microsoft have a huge article on it, but it does trip people up a lot.
The Azure Management Portal should also be able to give you a quick summary of if your roles & instances are actually up, healthy and stable.