7 Monitoring and Aanlytics

Amazon Cloudwatch

Monitor applications and infrastructure in real time
tracking and monitoring metrics
Metrics
- variables that are tied to your resources
Can apply rules around metrics (ie when reach a number)
- Can alert monitoring users (people or apps) to perform some action that will turn off the alert
- Cloud watch alarm (The rules to trigger alarm and action in Cloudwatch)
  - Integrated with SNS, so a notification can be sent to another service to perform an action
Dashboard
- a feature that allows a user to view the metrics in real time
- auto refreshing
Beneifts
- access to metrics from all places in a central locations
- visibility into your apps, infras and services
- system wide view
- Reduce MTTR (mean time to resolution) and improve TCO (total cost of ownership)
  - Frees up developers from fixing or spending time fixing issues and focus on features
  - Thus reducing costs and time
- Drive insights and optimise apps and resources

Trust but verify
Everything is recorded (auditing)
Auditing tool for recording api calls
- everything in AWS is an api call
Every request is recorded in cloud trail
Save logs in secure S3 buckets
Includes
- api called
- time of call
- source ip address
- how it was made
- etc
Events are updated in cloud trail withing 15minutes
- use of SNS/SQS or some streaming system like kafka
Cloudtrail insights
- optional
- automatically detects anomalies in api activities

Automated advisors
evaluate your resources against the following pillars:
- cost optimization
- performance
- security
- fault tolerance
- service limits
Some are free
can see in console
- shows three areas for each pillar
  - green: all good
  - orange: cost optimization
  - red: action recommended now