Slide 1

Slide 1 text

Topology Visualization for your Service With Application Map With one click enablement on Azure Alex Bulankou @bulankou

Slide 2

Slide 2 text

Microsoft Application Insights • Application Insights is an extensible APM service for web service developers on multiple platforms • We have SDKs for .NET, ASP.NET Core, JavaScript, Node.JS, Java, Python, PHP, Go • We are “Azure first but not only”: premium integration and enablement for Azure- hosted services, but support on-prem services and other cloud platforms • As long as you are willing to have your telemetry sent to Azure

Slide 3

Slide 3 text

Topology Diagrams is Often Best Way to Communicate (But can be a lot of work) • Great communication tool • Easily go out of date • Separated from performance/health data for each service

Slide 4

Slide 4 text

DEMO: Generate your Topology in One Click (1) Central node represents your service (2) Client node represents the clients onboarded to AppInsights JS SDK (3) HTTP calls from client are AJAX requests (4) 100% of failures in this case clearly identifies configuration error

Slide 5

Slide 5 text

DEMO: Diagnose Configuration Issue

Slide 6

Slide 6 text

Application Map for our Realistic Service We identify Azure worker roles, Azure queue, blob, storage and are adding identification for more dependency types (DocumentDB, Redis, EventHub) Availability tests results Custom thresholds

Slide 7

Slide 7 text

Adding Custom Dependency Tracking DependencyTelemetry customDependency = new DependencyTelemetry(); // set type, target (server name) customDependency.Type = "MySQL"; customDependency.Target = "mysql.net"; // data is full statement customDependency.Data = "select * from mytable"; customDependency.Duration = TimeSpan.FromMilliseconds(200); customDependency.Name = "mydb | command"; customDependency.ResultCode = "0"; customDependency.Success = true; telemetryClient.TrackDependency(customDepen dency); Dependencies appear grouped by type And then by target

Slide 8

Slide 8 text

How we run Daily Live Site Triage (2) Continue with metrics charts, to identify on call engineer to take investigation (1) Start with Application Map (3) After that investigation continues in Analytics

Slide 9

Slide 9 text

Coming soon: Application Map showing multiple components Service instrumented with AppInsights calls into another service instrumented with AppInsights • Supported by .NET and Node.JS SDK • UI is currently being flighted 25% • Works by injecting special headers into your request and response to detect that component in question is onboarded to Application Insights

Slide 10

Slide 10 text

North Star: combining auto-discovered and custom nodes Splunk glass table: you can add your nodes and associated metrics for each node. Then you are able to monitor the whole suite of micro-services with single pane of glass

Slide 11

Slide 11 text

Thank You!