EzCloudInfo

In the previous post we looked on how to configure Azure Monitor Alerts for Critical events that occurs on Microsoft Windows Devices which can be used for monitoring the Teams Room Systems. With Azure Log Analytics we could leverage few more components that will help us to visualize the status of the systems which are monitored through selected event logs and the performance counters.

Creating the Workbooks and making them visualize purely depends on the data that is been ingested on the corresponding log analytics workspace. So at the first stage its very important that we are sending all the required logs and counters which is mandatory for visualizing the metrics.

Firstly before creating the workbooks we need to devise a strategy on how to build a skeleton for the dashboard. This is very important since there are multiple options available and need to understand what important data that needs to be projected on the dashboard.

We will go through few examples of how to get started with creating the workbooks and visualizing the data.

We need to prepare the required Kusto Query Language which is required for visualizing the data. Below is a small example of one which will visualize the count of the perf counters by object name

Perf
| where TimeGenerated > ago(1h)
| summarize count() by ObjectName

To Render them as a pie chart we can use the below information

Perf
| where TimeGenerated > ago(1h)
| summarize count() by ObjectName
| render piechart 

Example below will project only the affected systems which has failed windows updates, driver updates or any devices connected with room systems which are in a failed state.

search *
| where Type == "Event" 
| where EventLog == "System"
| where EventLevelName == "Error"
| extend Status = parse_json(RenderedDescription).Description
| where RenderedDescription has "failed"
| project TimeGenerated, Computer , RenderedDescription 

If we need to visualize them on a graphical pie chart we could do that as well by summarizing them to a string value which is available from the  parsed json file. Example it can be computer, Ip address , Device name or any data which is present on the raw event data.

search *
| where Type == "Event"
| where EventLevelName == "Error"
| extend Status = parse_json(RenderedDescription).Description
| project TimeGenerated, Computer,RenderedDescription 
| where RenderedDescription has "failed"
| summarize Count=count() by tostring(Computer) 
| render piechart

Above are just very few examples of rendering the data and making them visualize through kusto query language. There is a lot to explore and can project more data based on the logs that we are adding on the azure log analytics.

Now we have got some idea of how to create the visualization through the kusto query language there is an option to combine multiple queries and display them as a dashboard through Azure Workbooks. Earlier this option was enabled by view designer which is now replaced by enhanced version called Azure Workbooks.

There are multiple options which can be utilized and created dashboards with Azure Workbooks and below we will go through few of the options which will help us in creating our customized workbooks.

In order to get started with Workbook – Navigate to the log analytics workspace – Choose Workbooks

Click on New

We get the default summary of our query from our workspace with the below piechart view.

If we want to go with our own query we can remove the default query and select Add. Here in Add we have multiple options like below out of which Add Group seems to be very much interesting. With Add group we have the ability to add multiple queries and group them in a single workbook.

At the top of this group we have an option to add text which visualizes the workbook name and the details.

After selecting the group , now we have option to add query into the group.

When going into the advanced settings we have these options now to display the chart titles specific for this query.

In the style tab we do have some options to modify the HTML settings. By default this will fit in to one query per row and if we need to add three queries we need to adjust the width settings.
In below case I have added the width to 50 since trying to add 2 queries in a row. But its very important to note here that adding 3 columns and making them visible as a dashboard is fine only in Azure Dashboards. If we try to view them from Azure Workbooks 3 queries in a row is not sufficient to accommodate and we do not have option to modify the HTML editor at this moment.

Have added another query which will let us know the status of the systems which have reported the heartbeat in last 30 minutes through the perfom counters. In below case since I have only one system for demo it shows only 1 system.

The moment when we group them and display it shows the  view as below. By adding multiple queries based on our requirement it makes us easier to create the dashboards.

Further to this we have a lot of options in the visualization of the display based on the metric units. We can go through few of them.
For instance there are below options available to set visualization.

We can reduce the size of the visualization and we have 5 options.

Further in the chart settings we have option to define the column and the units.

In the series we have option to change the colour and add a custom display label.

To interpret further have chosen Graph which is very interesting.When entering into the graph settings we have the below options in the node format settings. This helps us to choose what fields that we can display on the view of these images in the dashboard.

We have furthermore tweaking information on the layout settings. The hive clusters are looking really nice like honeycomb in the visualization. And there is a category to group by field to select based on the available fields.

Now we have the category to choose based on the coloring type. Ideally this is very good to categorize based on healthy and unhealthy systems. This will group the healthy and unhealthy systems separately and finally display them as dashboards.

This blog gives an overview of how to visualize , group and create Azure Workbooks from Log Analytics WorkSpace. With Azure log analytics and Azure Workbooks it makes very much easier to monitor the modern Windows 10 & Linux devices. This facility can be very much leveraged easily in a direct cloud deployment model without the need of installing, configuring and maintaining a local monitoring solution.

Thanks & Regards

Sathish Veerapandian

In the previous post we had an overview of how to create Azure Log Analytics and configure them to collect data from windows systems. Once the information is ingested in the workspace we currently have a choice to make alarms and notify the responsible team dependent on various signal logics which will be useful on monitoring these devices.

These alerts are scoped to each log analytics workspace. It will be a smart thought to isolate the services ,group them on singular workspace and create separate alerts for critical events happening on these monitored devices.

In order to create the alerts Navigate to alerts on the same workspace  – Click on New Alert Rule

Navigate to signal logic and choose the signal logic. There are multiple we need to see if any more interesting which suits our requirement can be added over here.

Now we have the required critical signals based on which the alert needs to be triggered. Usually the signal type will be from the collected events and the performance counters. In our scenario we could go with some default events from the list and also custom log search.

Device Restart Alert:

In our example for default one did choose the Choose the signal logic of heartbeat from the existing one – (Useful when the device turns off)

Select the required devices  – make operator threshold value 0 – aggregation 5 minutes & frequency of evaluation 1 minute (The frequency of aggregation and evaluation can be chosen based on the interval of how many times we want to check the heartbeat). In normal cases it is best recommended not to choose a smaller frequency time range for large volume of devices and probably for critical devices alone it can be selected on a smaller frequency time period.

Disk Threshold Alert:

Similarly like device restart we are having disk threshold alert by default which can be configured.

It notifies when it exceeds the configured space. Select the resource configured for Teams – Select the Condition – Select the computers – the object name whenever the % free space is greater than and choose the value 60 percent. The percentage can be altered based on our requirement.

Then we need to select the required object, instance , counter path and source system. In our case we have selected one performance counter % free space. This will alert us when the disk space crosses 60 percent of overall capacity.

Chosen aggregate period is 5 minutes and the frequency time is 1 minute for every evaluation. Again we can change the frequency of evaluation for this probably on two times in a day one on the earlier time and other one  on the evening.

Custom Alerts:

Custom Alerts are more intriguing. With custom alerts we must be able to avail most of our alerting mechanisms. We have to select the signal custom log search for the custom alerts.

Event  | where EventLog == "System" | where EventLevelName == "Error"
|where RenderedDescription != "*updatefailed*" 
| where EventData != "DCOM"
| project TimeGenerated, Computer, RenderedDescription

Example used the above query to report only the events which has error messages apart from windows update and DCOM alerts . We can further filter on not contains operator and create custom query based on  our requirement.

When any error messages apart from the excluded events comes up in the targeted devices we will be alerted for the same.

Note there are multiple action types – Email/SMS/Push/Voice, ITSM and Webhook will be more convenient for us in this case on Skype room systems monitoring.

Email – We can send email Email/SMS/Push/Voice when the alert is triggered. This will be the most convenient and easiest part to start with. This will help us to collect all the used cases initially and see which ones are really helpful and the ones which is not helping us. Once we devise a strategy from the email alerts then probably we can go with the other alerting mechanisms.

ITSM – We can integrate with IT service desk management tool to create incidents when these alerts are triggered. Most of the IT service desk management tools are capable of API integration especially with Azure AD and must be easier to suffice this requirement.

Webhook- We can configure to send push notification to teams channels when these alerts are triggered. Probably a dedicated teams channel can be created for the first level of NOC monitoring team. Post that the webhook can be configured to trigger the critical events alert to the teams channel.

Now with the email alert – Created action group – Chosen action type email/SMS/Push/Voice

By default there are no action group created. So an action group must be created and targeted to NOC team email group.

Added the email address for notification. Well there are other options as well like sending SMS and Voice which could also be leveraged.

We do have an option to modify the email subject based on the alert details.

Finally we name the alert details , mark the severity , enable and create them.

We have the option to see all the  configured rules.

Once after configuration, we can see the statistical dashboards which provides us the summary of total alerts that have been triggered and their status.

We are receiving the email alerts when the disk space exceeds the configured level of 60 percentage.

Similarly when the device was turned off, the configured heartbeat alert triggered an email to the recipient.

Similar like this we can create multiple required alerts for critical events.

At this moment we have option to create alerts for every action type which can be targeted for all computers and they are charged individually on a very nominal price. So for multiple alerting types we need to create multiple action types. These alerts are purely based only on the collected logs which are present on the azure log analytics workspace. Just in case if we are trying to collect more details which are not present on the collected logs then we wouldn’t be able to create the alerts. The Azure logs Alerting mechanisms provide a great way to alert the critical events happening across the monitoring systems.

Thanks in Advance

Sathish Veerapandian