Azure Machine Learning with Log Analytics
1 on OMS with Azure ML. As organizations lack the ability to mine through mounds of log data to detect trends and determine what services are running well and which services may need help.
Problem Management is one of the processes that requires Incident/Event data to determine trends. Many organizations struggle with this as they need to have a person look at this data and analyze it. Using OMS Analytics data and pumping to through Azure Machine Learning and providing insights would be valuable to IT organizations.
This would increase IT Organization’s maturity and drive business value.. I see a need for this and no one is doing this yet. Making IT organizations more efficient and leveraging data driven decisions is where we can gain momentum with OMS.
If someone in the Product Group needs me to define the User Story for this scenario, I would gladly assist.
The value prop for this integration of OMS and Azure ML seems pretty clear. Great bang-for-the-buck and goes well beyond just the telemetry based OMS solutions my team is working on.
Thank you for your feedback. This is supported today, please check the doc here:
Please feel free to revert back if you have any other feedback or existing product functionality is not satisfying your scenario. Thanks
Cathy Du commented
This is supported today, check the doc here:
Are there any updates regarding this integration between OMS and Machine Learning ?
Customer has very big deployment of GENESIS (SCADA) http://iconics.com used for monitoring data from 100k+ non-intellectual sensors and controllers.
He’s happy with GENESIS from basic monitoring perspective, but want to improve with advanced features:
• Advanced monitoring scenarios
• Data correlation with other 3rd party monitoring systems
• Visualizations with simple trend logic with Power BI for MIR Center
• Predictive analytics (ML)
• And other small features
Good part: GENESIS uses SQL as database with quite simple data structure.
Bad part: today I can’t find a way to push data from OMS to ML and this missing part is very important for Customer in nearest future.
How custom applications (.NET\WPF\WCF bases applications) behave after monthly patching (specially after installing the .NET and VC++ updates)?
While working on .NET setup cases customer often used to ask same question to support team.
I think we can run a check to identify certain event ids if they appeared while pervious patching.
“Just like Joseph I'm also engaged at helping a large bank move to the cloud and customer has indicated that being in control is super important for them while moving to the cloud.
That's why we are leveraging OMS, Azure Activity Logs, Automation, PowerBi etc.
By using Azure ML customers would be better equipped to even lower the risk by having better insight. With the insights of Azure ML, customers would be better equipped to assess the risks and be more in control.”
Joseph Marino commented
We are currently engaged in a large digital transformation / Azure automation project with a large bank in Australia. We are building a completely new operating model for the bank based on leveraging the PaaS services in Azure and VSTS. We have built some critical pipelines that automates much of the traditional Service Management processes.
Two key areas that would be assisted with the application of ML would be the Service Management processes that we cannot fully automate, which are Incident and Problem Management.
We are leveraging OMS and Azure Advisor in the solution and we have also built a compliance engine that closes the gap on the current Azure Compliance capabilities.
All of this we are looking to surface through PowerBI Dashboards and Drill Downs– however, we would love to see what could be done with ML and Log Analytics - specifically targeting Incident and Problem Management.
Please feel free to contact me for the user story or if there are any questions. Thank you.
Giulio Astori commented
Huge quantities of log data generated by all sorts of devices opens immense potential for insight, but machine learning is needed to make sense of it.
Rob Rinear commented
The capability of mass data collection is only as useful as the customer’s ability to query it and recognize value. If Azure ML can mine the data and provide customers with valuable insights that they otherwise could not glean themselves – patterns, anomalies, issues, etc. – then Log Analytics has an entirely new value proposition. Ideally there would be a configuration option within Log Analytics to indicate what data you want Azure ML to assess and have any findings or suggestions surface back in Log Analytics. It has to be "check-box" simple.