Connecting the dots with OMS and SLACK

In my last blog post, I spent some time trying to explain why OMS is more than you think and how this fits into the next generation of hybrid IT management, with Management-as-a-Service.

Why OMS is more than you think

Today, I want to highlight something I myself find very interesting, where we are using OMS as the source of our information towards operations engineers.

OMS Log Analytics

One of the key aspects of OMS is the Log Analytic workspace. This is where you harvest the data from your hybrid operational environment, and as I talked about in my previous blog post, you can have multiple data sources – and even use custom logs to retrieve and centralize the information you are looking for – but also (and perhaps more important) – the information that you didn’t knew you were looking for!

Log Analytics let you easily search for any of your data and from there, you can truly demonstrate your skillset by connecting the dots to a complete remediation solution, or plug into some other system to either deliver or manipulate the data or both.

With Log Analytics, we are able to:

·         Search for any of our data

·         Save searches and use them together with Dashboards

·         Use saved searches in conjunction with Alerts

·         Get e-mail notification with detailed information about the alert, search result and more

·         Connect Alerts with Azure Automation to trigger a Runbook that is either executed in Azure or through a Hybrid Worker

·         Connect Alerts with third-party systems using WebHooks

This blog post will focus on how to use OMS as the foundation for an operational department and centralize the alerts (informational, warning and critical) into SLACK.

First, let us quickly get a better understanding of what SLACK really is and why it might be useful in this particular scenario.

Many IT organizations are having a wide diversity of different ways of doing collaboration. Some of them are good, some of them are less good. However, the fact is that many channels might be used over time which leads to lack of communication and especially transparency when it comes to critical information around the operations.

Slack is a messaging application where teams can share files, talk and literally work together. This lets organizations have everything in one place, moving away from the devastating e-mail threads and so on.

With SLACK, everything that is shared is automatically indexed and archived which is searchable.

Some of the advantages you will get immediately when using SLACK is transparency to team communication for greater visibility into what other teams are working on, it speeds up feedback and decision making and make it a lot easier to find information and documents and more.

Las but not least – SLACK supports a wide range of tools, which means you can integrate existing apps, systems etc to communicate with SLACK to centralize the communication and information.

This is where OMS comes into play together with the WebHook integration to SLACK.

Ok, I get it. The information from our alerts can have a flow into one or more SLACK channels where our teams can get everything in a single view, but what exactly is a WebHook?

I am glad you asked.

WebHook is something you have used already if you have been using Azure Automation – and especially together with Alerts in OMS. This will leverage WebHooks.

The concept of WebHooks is really simple, and by simple I mean it is a simple HTTP POST that occurs when ‘something’ happens.

Using OMS together with SLACK, OMS will POST a message to a URL when certain things happen (the Log Analytic Search is showing some result that will trigger the Alert workflow).

WebHooks helps us to receive valuable information when it happens – instead of constantly pulling for the data.

In SLACK, you can add an ‘incoming webhook’ to your channel that will accept data from external sources that will send a JSON payload over HTTP.

Each channel in SLACK will be identified by a unique incoming Webhook URL to which you can post the message from outside.

A typical JSON payload will look similar to this:

{

  "text": "This is some random text from Virtualization and some Coffee",

  "channel": "#virtualization",

  "username": "Kristian",

  "icon_emoji": ":KristianDancing"

}

Once you have added the incoming WebHook to your SLACK channel, you can take advantage of this when creating alerts in OMS.

Here's an overview of the workflow and architecture

Here’s an example on how to configure an Alert in OMS to use a Webhook

And this is an example on how it could look like in SLACK, where we have different channels for different teams, depending on their area of expertise, responsibility etc.

Happy integrating!

OMS: Automated VM protection using Azure Recovery Services, Automation and Log Analytics

It’s me again, blogging from a random hotel room.

Last time we had a look into the combination of OMS and Azure in order to ensure availability of a Windows Server container hosted on a virtual machine in Azure.

http://kristiannese.blogspot.no/2016/02/azure-and-oms-better-together.html

(The templates and scripts will be published soon).

This time, I would like to focus on another scenario I think is very useful in the combination of Azure and OMS.

Recently, the GA of Azure Recovery Services were announced under the ARM APIs, which I talked about in this post and covered some considerations to be aware of - http://kristiannese.blogspot.no/2016/01/azure-site-recovery-and-azure-resource.html

In this blog post, I would like to cover a common scenario that we will run into every now and then when customers wants to protect virtual machines automatically to Azure in a programmatic way.

Overview

More and more customers are looking into how they can leverage the Azure cloud today, and one of the low-hanging fruits are services that can easily be plugged into existing services on-premises that will easily enable hybrid cloud scenarios, such as Business Continuity and Disaster Recovery. These services can be harnessed directly from Azure but provides you with a more comprehensive solution when used in conjunction with the entire OMS suite – that includes these services as well.

In a couple of minutes, you can have your services and applications running healthy as ever in an Azure region once this has been configured, in case of a failover.

In this article I will not go into design principals around the recovery processes (we’ll save that one for later), but rather cover a scenario that automatically will take care of some heavy lifting for you.

Use case

Many organizations have Hyper-V running as their primary hypervisor on-premises today, powering test, dev and production virtual machines. Since Azure has been able to democratize disaster recovery with their recovery services, people are looking into how to take advantage of this in a streamlined and efficient way. My goal here is to show how you can onboard and enable protection for newly created virtual machines on a Hyper-V host that has been registered to your Recovery Services Vault in Azure, by combining events that are logged into Log Analytics in OMS, monitored by a saved search that also has an associated alert with remediation attached to it.

This will invoke a Runbook created in Azure Automation that will enable replication on newly created virtual machines on that particular Hyper-V host and replicate to Azure as the recovery site.

Breakdown of the workflow

·       A VM gets created on the Hyper-V host. Regardless of how it’s created, this will log a specific informational event in the Hyper-V-VMMS-Admin log on the host, EventID 13002 that “A new virtual machine was created”

·       The OMS agent deployed on the host will fetch this and ingest this into Log Analytics

·       A search query is defined to monitor for this specific event on this specific host

·       An alert is created and associated with this query, so an e-mail will be sent when this occur

·       A runbook is created in Azure Automation that will search for newly discovered virtual machines on the registered Hyper-V host in the Azure Recovery Vault, look for VMs that isn’t protected and enable protection for them.

·       This runbook is associated with the alert and is part of the remediation process

Getting started

Before we can enable this scenario, we have to have some prerequisites in place:

·       OMS Workspace

·       Azure Automation account

·       OMS agent installed on the applicable Hyper-V hosts

·       Azure Recovery Vault in ARM

·       Runbooks

Assuming you have all of the above except the runbooks, I’ll cover the creation of the Azure Recovery Vault and the Runbook and stitch everything together.

Creating Azure Recovery Vault with Azure Resource Manager

The following PowerShell cmdlets will enable Azure Recovery Vault in your subscription and go through the creation of the Vault:

Note: You must install and register the agent on your Hyper-V host manually in this process, before proceeding with the rest of the script.

# Login to Azure and select Azure subscription

Login-AzureRmAccount -Credential (get-credential -Credential kristian.nese@ledzeppelin.onmicrosoft.com)

Select-AzureRmSubscription –SubscriptionID $subID

# Check to see if your Azure subscription is enabled with ASR resource providers

Get-AzureRmResourceProvider -ProviderNamespace "Microsoft.RecoveryServices"

Get-AzureRmResourceProvider -ProviderNamespace "Microsoft.SiteRecovery"

# Register ASR Resource Provider

Register-AzureRmResourceProvider -ProviderNamespace "Microsoft.SiteRecovery"

Register-AzureRmProviderFeature -FeatureName "betaAccess" -ProviderNamespace "Microsoft.RecoveryServices"

Register-AzureRmResourceProvider -ProviderNamespace "Microsoft.RecoveryServices"

# Verify to see that the RP is registered correctly

Get-AzureRmResourceProvider -ProviderNamespace "Microsoft.RecoveryServices"

Get-AzureRmResourceProvider -ProviderNamespace "Microsoft.SiteRecovery"

# Create a new ASR Recovery Vault

mkdir -Path "c:\ASRDemo"

$RG = "DRGroup"

$Location = "west europe"

$vault = "HyperV"

$path = "c:\ASRDemo"

New-AzureRmResourceGroup -Name $rg -Location $location

$vault = New-AzureRmRecoveryServicesVault -Name $vault -ResourceGroupName $rg -Location $location

$FilePath = Get-AzureRmRecoveryServicesVaultSettingsFile -Vault $vault -Path $path | % FilePath # changed this to make sure we get the right filepath

Import-AzureRmSiteRecoveryVaultSettingsFile -Path $FilePath

# Create a new Hyper-V Site in the ASR vault

$sitename = "KNHVSite"

New-AzureRmSiteRecoverySite -Name $sitename

# Query the job

Get-AzureRmSiteRecoveryJob

# Generate and download registration key for the site (Copy the downloaded key to the Hyper-V host. You'll need the key to register the Hyper-V host to the site)

$SiteIdentifier = Get-AzureRmSiteRecoverySite -Name $sitename | Select -ExpandProperty SiteIdentifier

Get-AzureRmRecoveryServicesVaultSettingsFile -Vault $vault -SiteIdentifier $SiteIdentifier -SiteFriendlyName $sitename -Path $Path

# Download and install the agent on the Hyper-V host(s) from this URL: https://aka.ms/downloaddra

# Verify that the Hyper-V host is registered in ASR

Get-AzureRmSiteRecoveryServer

# Create storage account with Geo-redundant storage

$storageaccountID = New-AzureRmStorageAccount -ResourceGroupName $RG -Name "knasrdemo01" -Type Standard_GRS -Location $Location

# Create a replication policy and associate it with the protection container

$ReplicationFrequencyInSeconds = "300"

$PolicyName = “DefaultPolicy”

$Recoverypoints = 1

$storageaccountID = Get-AzureRmStorageAccount -Name "knasrdemo01" -ResourceGroupName $RG | Select -ExpandProperty Id

$PolicyResult = New-AzureRmSiteRecoveryPolicy -Name $PolicyName -ReplicationProvider “HyperVReplicaAzure” -ReplicationFrequencyInSeconds $ReplicationFrequencyInSeconds  -RecoveryPoints $Recoverypoints -ApplicationConsistentSnapshotFrequencyInHours 1 -RecoveryAzureStorageAccountId $storageaccountID

# Get the protection container and start the association

$protectionContainer = Get-AzureRmSiteRecoveryProtectionContainer

$Policy = Get-AzureRmSiteRecoveryPolicy -FriendlyName $PolicyName

$associationJob  = Start-AzureRmSiteRecoveryPolicyAssociationJob -Policy $Policy -PrimaryProtectionContainer $protectionContainer

Once this has been created and the Hyper-V host(s) are registered to Azure, you should be able to see the virtual machines on that host by executing the following cmdlet:

Get-AzureRmSiteRecoveryProtectionEntity -ProtectionContainer $protectionContainer | select friendlyname

FriendlyName

------------

ASRGEN2-01     

ASRGEN2-03 

ASRGEN2-02  

Preparing Azure Automation

There’s a few things we need in order to successfully create and run our runbook in Azure Automation

·       Assets

o   Credentials

o   Variables

o   Modules

·       Runbook

For the assets, you can easily add the required assets by using PowerShell.

In my case, I need the credentials to login to an Azure subscription, and I also need to store the subscription ID in a variable.

This can be achieved using the following cmdlets:

Creating Automation Variable

New-AzureRmAutomationVariable -Name SubscriptionID -Encrypted $false -Value $Subscription -ResourceGroupName $RGName -AutomationAccountName $AAName

Creating Automation Credential

New-AzureRmAutomationCredential -Name AutoAdmin -ResourceGroupName $RGName -AutomationAccountName $AAName -Value $cred 

Uploading Automation Module

Normally you would use a cmdlet for uploading PowerShell modules to Azure Automation as well, but since we just want to grab a few of them directly from the available PowerShell Gallery in the AA itself, we quickly head over to the portal and grab them from there.

Install the following modules in this specific order.

Note: it can take several minutes before the modules are installed and ready. You might find yourself a cup of coffee while doing this, as some of these modules has dependencies of each other and won’t import before the dependencies has completed the import process

Creating the Runbook

Now that the prereqs are in place, it is time to author the Runbook that will do the following:

Runbook

# Runbook for enabling DR on unprotected Hyper-V VMs

$Admin = Get-AutomationPSCredential -Name AzureAdmin

$Subscription = Get-AutomationVariable -Name SubscriptionID

Login-AzureRmAccount -Credential $Admin

Select-AzureRmSubscription -SubscriptionId $Subscription

# Fetching some variables for this particular setup

$RG = "DRGroup"

$Location = "west europe"

$vault = "HyperV"

$sitename = "KNHVSite"

$vault = Get-AzureRmRecoveryServicesVault -Name $vault -ResourceGroupName $rg

$storageaccount = "knasrdemo01"

# Creating a temp directory to store the VaultSettingsFile

$tempFileName = $env:TEMP

$FilePath = Get-AzureRmRecoveryServicesVaultSettingsFile -Vault $vault -Path $tempFileName | % FilePath # changed this to make sure we get the right filepath

Import-AzureRmSiteRecoveryVaultSettingsFile -Path $FilePath

$SiteIdentifier = Get-AzureRmSiteRecoverySite -Name $sitename | Select -ExpandProperty SiteIdentifier

Get-AzureRmRecoveryServicesVaultSettingsFile -Vault $vault -SiteIdentifier $SiteIdentifier -SiteFriendlyName $sitename -Path $tempFileName

$protectionContainer = Get-AzureRmSiteRecoveryProtectionContainer

# Get Storage Account

$storageaccountID = Get-AzureRmStorageAccount -Name $storageaccount -ResourceGroupName $RG | Select -ExpandProperty Id

# Get Recovery Policy

$policy = Get-AzureRmSiteRecoveryPolicy

# Enabling VMs for protection

$protectionEntity = Get-AzureRmSiteRecoveryProtectionEntity -ProtectionContainer $protectionContainer | Where-Object {$_.FriendlyName -like "*asr*" -and $_.ProtectionStatus -eq "Unprotected"}

foreach ($entity in $protectionEntity)

     {

        if ($protectionEntity.ProtectionStatus -eq "Unprotected")

        {

            Write-Output "They aren't protected!"

            Set-AzureRmSiteRecoveryProtectionEntity -ProtectionEntity $entity -Policy $Policy -Protection Enable -RecoveryAzureStorageAccountId $storageaccountID -OS Windows -OSDiskName $protectionEntity.Disks[0].Name -Verbose

        }

        else

        {

            Write-Output "They were protected already :-)"

        }

    }

If you want to use this in your environment, ensure you are changing the variables to meet your needs.

OMS

Assuming you already have the host registered to the Workspace, you should do the following steps to be able to get the information you are looking for to enable this scenario where you want to leverage the runbook to protect newly created VMs.

Adding logs to OMS

In the OMS workspace, Click on ‘Settings’

Navigate to ‘Data’ and click on ‘Windows Event Logs’

Add the following log that will contain the information about creation of virtual machines on the host

Ensure that Error, Warning and Information is selected

Note: OMS doesn’t care about what has already happened, so only new events in this log will appear in OMS.

Next, go back and drill into ‘Log Search’

I use the following search to pinpoint the specific EventID and the Hyper-V host

EventID=13002 host11 | Select Computer, Activity, TimeGenerated, Message, EventLog

I have saved the search and categorized it as ‘Hyper’V.

Now, I want to enable alerting on this search, so I click on the search and then on the ‘Alert’ button

Assign a name to the alert and specify when the query should run and when an alert should be generated.

I then specify the recipient of the alert and give it a name.

As a last step, I connect the alert with the newly created Runbook in Azure Automation and ensure that it will be executed by using an Azure worker and click save.

Creating a new VM to trigger the alert

Heading over to my Hyper-V host I created some new virtual machines.

Coffee time

Since OMS will use this search query every 15 mins, I had enough time to make myself some coffee while waiting for the e-mail to drop in my inbox

Alert

Once the search query detect a new event, the alert is triggered and an e-mail is fired away to my inbox.

Remediation

This should invoke the associated runbook for remediation, and when I check into the Job view in Azure, I can see that it has successfully been executed and also the output that tells me that replication has now been enabled on the unprotected VMs.

Verifying the remediation

Using PowerShell, I can access my Recovery Vault and check to see which VMs are in the process of being protected in Azure by using the following cmdlet:

Get-AzureRmSiteRecoveryProtectionEntity -ProtectionContainer $protectionContainer | select FriendlyName, ProtectionStateDescription

The output shows that several machines are already protected while the newly created VM is in the process of doing an initial replication

FriendlyName ProtectionStateDescription       

------------ --------------------------       

ASRGEN2-04   Protected                        

ASRGEN2-02   Protected                        

ASRGEN2-01   Initial replication is in progress

ASRGEN2-03   Protected                        

Closing note

This was yet another blog post that shows some of the capabilities of combining OMS, Azure and Azure Services together.

Moving forward, I will focus more on some of these scenarios to show where we have some gaps today, and how you can overcome those gaps with a little effort in engineering.

By the end of the day, we can more or less achieve whatever we want as long as there’s an API for it, and that’s exactly what this blog post was showing.

Thanks!