Application insights overview

Azure API Management

Fiodar Sazanavets

Senior Software Engineer at Microsoft

Importance of observability

Observability analogy

Telemetry types

Logging
Tracing
Metrics

Introduction to logging

Logging records each event in human-readable language
Every log entry has a severity level (normal, warning, error)
- Like a secretary taking notes during a meeting
Captures everything happening inside the application

Logging

Overview of tracing

Traces

Multiple related log entries can form a trace
Imagine each request as a parcel with a tracking sticker
APIM attaches a unique correlation ID into request headers
Backends with Application Insights SDK detect the ID automatically
This lets the platform stitch together the full journey of the request

Importance of metrics

Metrics are simple measurements (counters, averages, durations)
- Like a car's speedometer showing current speed
Examples of metrics:
- Error counts
- Average request latency
- Number of active users
Always tracked as a time series to show what was happening when

Metrics

Application insights dashboards

Dashboard

SLAs = speed limits
- The siren goes off when thresholds passed

Dashboards = your wall of monitors
- One glance should answer: Are we fast? Are we failing?
- In Azure Monitor Workbooks/Dashboard, pin charts for request count, error rate, and latency
Alerts = your seatbelts
- Create alert rules tied to charts or logs
- Wire to Action Groups (email, Teams, PagerDuty)

Kusto Query Language

When something smells off, KQL is your detective
Open Logs to query application telemetry
- Like flipping through a flight data recorder
Saved queries let you jump from alert fired -> root cause in minutes
Confirm the fix by watching gauges settle back into green

KQL example

Let's practice!

Azure API Management

Preparing Video For Download...