Azure

What are the conventions for queues, tables, and blob containers? Unfortunately I can’t find any recommendations anywhere and they are not shown on this page either? https://docs.microsoft.com/en-us/azure/cloud-adoption-framework/ready/azure-best-practices/resource-naming

All I can find is the convention for the storage accounts themselves (stnavigatordata001).

I can think of the following:

• Queues: st-queue-whatever-001, queue-whatever-001, q-whatever-001
• Tables: st-table-whatever-001, table-whatever-001
• Blob containers: st-blob-container-whatever-001

Hi all,

I need to call an API and copy the data as text files into a blob landing area. Because of the network security concerns, we thought of using managed private endpoints.

Here’s the surprising part we didn’t expect. Each instance of copy activity takes 2 to 3 minute to warm up. It’s not even at a data factory level or a pipeline level. If you had your copy activity in an until loop with 10 iterations(pagination on our source API), your pipeline that runs in less than 2 mins in normal IR takes more than 20 to 30 minutes when using managed Vnet. This is just for one source table API, if we were to do this for like 50 or more tables a day, it obviously becomes unusable or your pipelienes become unnecessarily long-running (2:45s queue. 0:01s copy duration).I am not sure what I can do in this scenario? Can I reach out to Blob storage using a self-hosted IR?I am guessing a moderately sized VM as a SHIR may be better than every copy activity in your ADF queueing for 2 to 3 minutes every time it runs. Have you faced this issue before? What would be a good way to approach this. From the security restrictions standpoint, we cannot expose the storage account to public internet or enable “Trust all Azure services” option.

Please help.

MS doc ref:https://docs.microsoft.com/en-us/azure/data-factory/managed-virtual-network-private-endpoint#activity-execution-time-using-managed-virtual-networkPS: We also raised a support ticket with MS and they basically said the same thing as the docs.Is this also the case for managed private endpoints for Azure SQL also? Does each hit get queued for like 2 – 3 mins? How is this a good thing? I am very curious in what cases this might be a good thing

I’m very good with Terraform and I’m trying to do some work with Bicep.

I understand Bicep is based on ARM, and I’m familiar with the MS docs that describe their templates.

However, I’m having a difficult time because the documentation doesn’t seem to provide details about the values that are expected or accepted for the various properties.

Say I want to create an Azure Virtual WAN. This is the template reference.

I need to define the following:

resource symbolicname 'Microsoft.Network/virtualWans@2020-07-01' = { name: 'string' location: 'string' tags: {} properties: { disableVpnEncryption: bool allowBranchToBranchTraffic: bool allowVnetToVnetTraffic: bool office365LocalBreakoutCategory: 'string' type: 'string' } } 

This is a simple example and most properties are mostly self-explanatory. But what about the properties that aren’t self explanatory, like “type”?

The table on that page linked above says:

Where do I find a list of acceptable values to provide for “type”? I get that I need to provide “The type of VirtualWAN”… but wtf are my options?

With Terraform, the docs tell me exactly what options are available – it’s “Basic” or “Standard”.

What am I missing? Where is that documented for authors of ARM/ Bicep? I shouldn’t have to go to the Terraform docs lol.

Just to be clear – I’m not asking for someone to tell me “oh, vwan types are X, Y, and Z”. I want to know where that’s documented, I’ve got a lot of Bicep to write..

Background: I’ve got a web app that I deployed from the vendor’s AMI but it’s basically just a Windows VM with their software pre-installed. It hosts pre-packaged applications that users can then download and run locally without needing to do a full install. Some of these can be really big (20gb+) and while there’s a client piece, it’s really just pulling them down over HTTPS.

So I noticed that downloads from this VM were feeling really sluggish. Based on the time I was only getting about 80mbps on our corporate network and 30mbps from home on gigabit fiber. Went through some of the Azure troubleshooting docs and did the iperf tests they recommended, and with 32 streams I was pushing 1gbps, which seems about right for the instance size, but with a single stream getting the same poor performance as before. Spun up another VM in the same vnet/subnet, slightly bigger instance just to test, and same iperf results. I also ran the download test over at https://www.azurespeed.com/Azure/Download and got the same numbers.

I’m a n00b when it comes to Azure but can it really be that slow? Is there some other better solution or product for this scenario I just don’t know about? Most of my search results so far are leading me down the path of custom web apps but I don’t have much control over the application itself. I did open a case with support but it feels like they are trying to wash their hands of it pretty quickly because ‘public internet’. Any help/suggestions would be greatly appreciated.