Automate and Send Alerts for Elastic Jobs Failures

by Contributed | Dec 12, 2020 | Technology

This article is contributed. See the original author and article here.

Alerts can be sent for any particular state of an elastic job. In this article, we discuss the case of ‘Failed’ jobs, however the solution can be easily extended to other scenarios. The script can be executed on client machines in background on a schedule or in Azure Automation Runbook. If failed jobs are identified, email notifications are sent, with an attachment of a log file containing only the failed jobs. 

Elastic Jobs are currently in public preview and the monitoring capabilities are resuming for now to ad-hoc executions of T-SQL / PowerShell commands to retrieve the elastic job execution status, with no embedded option for sending alerts in case of a job failure. 

This need has been addressed through a PowerShell script that filters job executions based on their status and sends the output to a log file, that is further sent as an attachment in a notification email and in the same time the log file is stored in an Azure Storage Account for further reference. 

script output when executed on client machine

The jobs are filtered with the below T-SQL, that can be modified as per the requirements:

SELECT * FROM jobs.job_executions WHERE lifecycle = 'Failed' ORDER BY start_time DESC

Other possible values for job execution states are Created, InProgress, WaitingForRetry, Succeeded, SucceededWithSkipped, Failed, TimedOut, Canceled, Skipped and WaitingForChildJobExecutions.

Generated log file:

log file snippet

Save log file to storage account:

# Get key to storage account
$acctKey = (Get-AzStorageAccountKey -Name generals -ResourceGroupName general).Value[0]

# Map to the reports BLOB context
$storageContext = New-AzStorageContext -StorageAccountName $StorageAccountName -StorageAccountKey $acctKey

# Copy the file to the storage account
Set-AzStorageBlobContent -File $output_file -Container StorageContainerName -BlobType "Block" -Context $storageContext -Verbose

save log to storage account

The email functionality can be leveraged through any Simple Mail Transfer Protocol server. The proposed script is using smtp.mail.yahoo.com on port 587. 

email notification alert

How to run and automate alerts

As elastic jobs can be scheduled to run on a regular basis from both PowerShell and T-SQL, so our alert script can. Depending on the desired deployment type you choose, there are several options. For any of the below options, you have to set the required parameters in the script: Insert SQL Server Details, Credentials & SMTP server details:

$server = '.database.windows.net' # Set the name of the server
$jobDB = ''   # Set the name of the job database you wish to test
$user = '' # Set the login username you wish to use
$passw = '' # Set the login password you wish to use
$FromEmail = 'xxxxxxxxxxxx@yahoo.com' # "from" field's account must match the smtp server domain 
$FromEmailPassw = '' # use app/server access token - it works with account passw
$ToEmail = 'xxxxxxxxxxxx@yyyyy.com'
$SMTPServer = "smtp.mail.yahoo.com" #insert SMTP server
$SMTPPort = "587" # or port 465
$StorageAccountName = ''
$StorageContainerName = ''

OPTION#1 

Run the script on schedule in background on client machine

In order to run it you need to:

• 
  

  Open Windows PowerShell ISE in Administrator mode

  
  


• 
  

  Open a New Script window

  
  


• 
  

  Paste the content in script window

  
  


• 
  

  Run it

  
  


• 
  

  If failed jobs are found, an alert email will be triggered and the log file containing details on the failed jobs will be attached to the email and either saved locally or sent to a storage account. The result of the script can be followed in the output window.

  
  


• 
  

  To run the script in background on a schedule, you can use the following commands:

  
  

##  =========================================================================
##                Schedule Commands for Client Machines 
##  =========================================================================
## To schedule script execution in background, please see below options

#insert parameter values
$script = 'script.ps1' # insert script path
$Time= 'MM/DD/YYYY HH:MM' # insert desired start time for schedule
$jobName = 'Job1'#insert desired job name


# display all scheduled jobs
Get-ScheduledJob

# add new job to run at schedule
Register-ScheduledJob -Name $jobName -FilePath $script -Trigger (New-JobTrigger -Once -At $Time `
    -RepetitionInterval (New-TimeSpan -Minutes 1) -RepetitionDuration ([TimeSpan]::MaxValue))

# command to remove a scheduled job
Unregister-ScheduledJob $jobName

OPTION#2

Run the script from Azure Runbook

• Create a new Automation Account as described here and make sure you choose “YES” for option “Create Azure Run As Account”.
  


• Import the following Azure Modules by browsing the gallery: 
          Az.Accounts (≥ 2.2.2)
          Az.Storage
          Az.Automation

             browse gallery – add module

• Create a runbook to run the script and make sure you choose Powershell runbook type.

           add a runbook

• Add the following login section when connecting from a Runbook        

##  =========================================================================
##             Login when running from a Runbook
##  =========================================================================
## Enable this code section when running from Azure Runbook
## Get the connection "AzureRunAsConnection" when run from automation account

$connection = Get-AutomationConnection -Name AzureRunAsConnection
Connect-AzAccount `
    -ServicePrincipal `
    -Tenant $connection.TenantID `
    -ApplicationId $connection.ApplicationID `
    -CertificateThumbprint $connection.CertificateThumbprint

"Login successful..."

• Schedule task

         schedule task

Note: Script execution can be monitored in the portal and alerts can be set for any script execution failure.

job execution monitor

The script can be found in the following Git repository. This project welcomes contributions and suggestions.

Build a virtual IoT Lab with Raspberry Pis and Azure IoT Hub

by Contributed | Dec 11, 2020 | Technology

This article is contributed. See the original author and article here.

Track the lineage of your organization’s data with Azure Purview

by Contributed | Dec 11, 2020 | Technology

This article is contributed. See the original author and article here.

Data lineage in Azure Purview helps organizations to understand the data supply chain, from raw data in hybrid data stores, to business insights in Power BI. Azure Purview’s turnkey integrations with Azure Data Factory, Power BI, Azure Data Share and other Azure Data Services automatically push lineage to Purview Data Map. 

Azure Purview also supports Apache Atlas Lineage APIs that can be used to access and update custom lineage in Purview Data Map. Hook & Bridge support from Apache Atlas can also be used to easily push lineage from the Hadoop ecosystem.

Figure 1: Data lineage can be collected from various data systems

Azure Purview can stitch lineage across on-prem, multi-cloud and other platforms

Enterprise data estate contains data systems performing extraction, transformation/load, reporting, ML (machine learning) and so on. The goal of lineage feature in Purview is to capture the data linkage at each data transformation to help answer technical and business questions.

For instance, Purview’s lineage functionality will help capture the data movement and transformation stages such as the one described below.

1. Data Factory would copy data from on-prem/raw zone to a landing zone in the cloud.


2. Data processing systems like Synapse, Databricks would process and transform data from landing zone to Curated zone(staging) using notebooks or job definition.


3. Data Warehouse systems then process the data from staging to dimensional models for optimal query performance and aggregation.

Data Analytics and reporting systems will consume the datasets and process through their meta model to create a BI (Business Intelligence) Dashboard, ML experiments etc

Root cause analysis scenarios

Azure Purview can help data asset owners troubleshoot a dataset or report containing incorrect data because of upstream issues. Data owners can use Azure Purview lineage as a central tool to understand upstream process failures and be informed about the reasons for discrepancies in their data sources.

Figure 2: Azure Purview lineage capability showing troubleshooting steps for a possible issue with Power BI report

Impact analysis scenarios

Data producers can use Azure Purview lineage to evaluate the downstream impact of changes made to their datasets. Lineage can be used as a central platform to know all the consumers of their datasets and understand the impact of any changes to their dependent datasets and reports. For instance, data engineers can evaluate the downstream impact for a deprecating column in a table or change in data type of a column. The data engineers can use Purview lineage to understand the number data assets potentially impacted by the schema changes of an upstream table. The column level lineage precisely points to the specific data assets that are impacted.

Figure 3: Azure Purview lineage capability showing the impact analysis for an upstream change

Lineage sources

Azure Purview can connect with Azure Data Factory, Azure Data Share, Power BI to collect lineage currently. In the coming months many more data systems such as Synapse Analytics, Teradata, SQL Server and so on will be able to connect with Azure Purview for lineage collection.

Call to Action

We are looking forward to hearing, how Azure Purview helped perform troubleshooting and impact analysis of your data pipelines with the native lineage experiences.

1. Create an Azure Purview account now and start understanding your data supply chain from raw data to business insights with free scanning for all your SQL Server on-premises and Power BI online


2. Start by connecting a Data Factory or Data Share account to push lineage.


3. Scan a Power BI tenant to see lineage in Purview. Use managed identity (MSI) authentication to set up a scan of a Power BI tenant


4. Learn more on lineage user guide.