Microsoft Azure’s CTO and Technical Fellow, Mark Russinovich, speaks about how the company helps keep up with demand for the newest technology innovations
“I’m excited to share the latest investments made in Project Flash, by the Azure Core Production Quality Engineering team. Earlier this year, we introduced Project Flash in an Advancing Reliability blog series to reaffirm our commitment to empowering Azure customers with monitoring VM availability in a robust and comprehensive manner. Today, we’re delivering holistic monitoring experiences that will meet customers’ distinct needs.”—Mark Russinovich, CTO, Azure.
For this multi-year project, we’re stoked to announce the Flash release – a completely new availability monitoring solution for Azure. We aim to provide customers with more accurate and comprehensive notifications about their virtual machines, enabling them to seamlessly deploy and scale anywhere.
Azure Resource Graph is a powerful monitoring platform designed for simplifying and enriching the way you monitor your resources in Azure, specifically VMs. It makes it easy to monitor and analyze VM availability, along with detailed downtime investigations and impact assessments.
Azure Monitor helps monitor and debug availability metrics, including the ability to view VM availability trends on a time-series view. It also lets you set threshold-based alerts when workload performance may be impacted, so you can quickly identify issues and troubleshoot them.
Azure Event Grid notifies you of VM availability status change events for instantaneous notifications on critical changes in VM availability. This allows Microsoft to quickly trigger remediation actions should any such action be required.
With Portal, it’s easy to monitor the health of your web server and understand your website or app’s performance. We’re committed to maintaining data consistency and similar rigorous quality standards across all the monitoring solutions that are included in Flash. Such solutions include existing responsibilities like Resource Health or Activity Log, so we deliver a consistent and cohesive experience to customers.
Azure Resource Graph provides a graph of your VM availability information in one space, with at-scale data for easy analysis.
Microsoft VMs running in Azure now have the ability to be tracked through a 14-day change tracking mechanism. With these changes due to general availability, VM availability information is now available when browsing resources in HealthResources.
Monitor Azure VM Availability with regular and fleet-wide VM snapshots without expensive querying times or high latencies.
The purpose of the Apollo platform is to provide accurate assessment of the impact to fleet-wide SLAs and rapid responses, in order to mitigate potential disruptions and predict failures.
What’s more important to you? Reliable alerts, controllers and benchmarks or a whole ‘lotta add-on features? If it’s the former, then you need ARG. As part of your server management framework, we process the information about your VM availability in conjunction with other metadata that is out there. We do this to give you complete visibility into how your system is performing.
With the VM change-tracking feature, you have access to a rolling 14-day window so that you’ll never miss changes in VM availability. You can also use this feature for detailed investigations.
Getting started
You can access your additional resource group data from Azure with either the PowerShell, REST API, or the Azure CLI. The following steps detail how data can be accessed from Azure Portal.
After installing the application you can access the Azure Portal by navigating to your settings. From there, navigate to Resource Graph Explorer and use it as shown below.
A portal view of Azure Resource Graph displays the list of datasets and the HealthResources table. Kusto queries can be run to access results in that table.
Figure 1: Azure Resource Graph Explorer landing page on Azure Portal.
You can retrieve the latest snapshot of VM availability information by selecting the Table tab and selecting the HealthResources table.
When looking at whether your resources are online, it can be tough to get an accurate picture. We created a tool to help you easily assess whether they are available by correlating view counts and annotations across all of them. To further the visualization, we pulled in event data related to HealthResources that you can use to find patterns in your infrastructure.
Figure 2: Azure Reserach Graph Explorer Window depicting the latest VM availability states and VM health annotations in the HealthResources table.
There will be a single column of events populated in the HealthResources table:
Azure Resource Graph is a new way to build your applications through a predictive database service. Portal view of the left-hand pane in Azure Resource Graph displays the 3 events within the HealthResources table along with the column used for each event type.
What’s your next step? You can learn more about Resource Graph Explorer by following these links.
resourcehealth.microsoft.com/availability-statuses
This event indicates the latest availability status of a VM, based on the health checks performed by the underlying Azure platform. Below are the current states we emit for VMs:
The tour is available and running as expected.
Unavailable: We’ve detected some form of abnormal activity in the virtual machine.
ScrollBack for sentence rewriter
This content is often generated by a platform that can’t accurately detect the health of a virtual machine. Users can usually check back in a few minutes for an updated status.
If you want to know the latest status of VMware, refer to the properties field which includes the below:
One-to-one correspondence equates to confidence in your work and success. With RHC, you can connect with business owners every day and share ideas.
targetResourceType
When you want to know more about health data flows, there are several types of literature that can help you.
resourceType
resources
Resource Id
resourceId
happenedTime
Check back for new availability updates
eventTimestamp
The ability to increase your reach by helping people find the correct thing at the correct time, and at just the right price, is accessible through Product Availability Simulator.
The previous state of the VM
the previousHealthStatus
availability
The current state of the VM is available.
currentHealthStatus is a program that gives you a way to do just that! You’ll have the ability to integrate your customer management system with our solution, and we’ll provide you with data on how well it’s working.
When building a survey, try starting out with general questions instead of focusing on a particular topic. The most common types of surveys ask about demographics, family history, general health and wellness, etc.
Microsoft.ResourceHealth/Resourceannotations
If a VM has failed, our new event contextualizes the disruption by providing necessary failure attributes in order to help users investigate or mitigate any change or disruption. We now emit the full list of previously non-existent properties on the VM health annotations that are required for investigation and correction.
These annotations are broadly divided into three categories:
When Amazon Aurora is down unexpectedly and the host crash-recovers you’ll get a downtime notification. The system will emit a virtual machine availability transition event, which is associated with each VM’s availability.
During control plane activities, Informational annotations are emitted by the system without any impact to VM availability. (For example, when an allocation, stop or delete is issued for a VM.) These messages do not require additional customer action, although you may choose to log and monitor them for potential issues.
VM degradation: This phenomenon is activated when VM availability is threatened. If you’re using a predictive failure model, for example, the VM could be at risk of crashing and losing data anytime. In that scenario, we’ve provided a deadline for when you should redeploy it for your best interests. Make sure to do so before then!
To get associated VM health annotations for a resource, refer to the properties field of the resource. There are two properties: InstanceHealth and InstanceHealthStatus. The InstanceHealth property includes all the current health information and attributes related to a particular instance, while the InstanceHealthStatus property specifies whether or not an instance is unhealthy or healthy, and also indicates if any fatal errors exist on it.
Corresponding RHC field
destination
Health data flows through a variety of channels, including a type of resource
It would seem that resourceType
fullTargetResourceId
Instead of rewriting the sentence, change the original word to something else that is not in the original.
Resource Id
resourceId
occurredTime
Platform generates a timestamp when the latest availability state is emitted.
eventTimestamp
MyAnnotation
This eventName is coming up!
reason
Now that you’re familiar with the overall availability impact, let’s talk about how specific hardware configurations will impact availability for your application.
The title
Description: Additionally, the listing is a tag that categorizes your listing. It signs up your post to be grouped with related posts, which means it’s more likely to show more often in search results.
The purpose of this field is to define the type of defect that triggered the annotation. This field can be used in conjunction with the type and source fields of this object.
Possible values: Planned | Unplanned | Not Applicable | Null
we love
Context
The Azure Availability Service includes a configurable activity indicator that enables you to track the reason behind any Azure Services or resources group changes. It is an inclusive method of tracking the availability status of your production environment and detecting any deviations before they become customers’to-notice.
Possible values: Platform-initiated | User-initiated | VM-initiated | Not Applicable | Null
Context
Here is the summed-up version of our content.
This is a statement about why annotation emissions are occurring, with remediation steps such as periodic browser scans for fingerprinting that can be taken by users.
The summary is a concise and well-organized overview of the article.
The public data for this topic is located here. This document covers some starter queries, but does not replace necessary research.
