CloudWatch

· Monitoring tool for AWS resources and applications

· Display metrics and create alarms that watch the metrics and send notifications and send notifications or automatically make changes to the resources you are monitoring when a threshold is breached.

· Cloudwatch does not aggregate data across regions. Therefore, metrics are completely separate between regions.

· CloudWatch Concepts

o Namespaces - a container for CloudWatch metrics.

§ There is no default namespace

§ The AWS namespace uses the following convention AWS\Service

o Metrics - represents a time-ordered set of data points that are published to CloudWatch

§ Provisioned metrics

v Basic - 5 minutes (default )

v Detailed - 1 minute (paid )

v CPU (if T2/3 can get credit usage for the Burst), Network IO , Disk Read/Write/Ops (instance store backed) and status checks (EC2 VM or System)

§ Custom Metrics (yours to push)

v Basic resolution - 1 minute

v High resolution - all the way to one second

v includes RAM, Swap, Application level metrics

v make sure the permissions on the EC2 role are correct (CloudWatchFullAccess)

o Dimensions - a name/value pair that uniquely identifies a metric.

o Statistics - metric data aggregations over specified periods of time

§

· Percentiles

· Alarms

o Alarms are used to Trigger notifications for any metric

o Alarms can go to Auto Scaling, EC2 instances, sns notifications

o Various Options (sampling, %, max, min etc ...)

o Alarm States

§ OK

§ Insufficient Data

§ ALARM

o Period

§ Length of time to evaluate the metric

§ High resolution metrics can be 10, 30 seconds

§ High resolution custom metrics can be 1+ second

· CloudWatch Dashboard

o multi region and can display any widget to any region, to do so just change to the region of the widget and add to the dashboard (then save)

· CloudWatch Events

o Source + Rule => Target

o Schedule: Cron jobs

o Event Pattern: Event rules to detect a service doing something,

o Triggers to lambda functions, SQS, SNS, Kinesis Messages

o CloudWatch event create a small JSON document to give information about the change

· CloudWatch Log

o Features

§ Monitor logs from EC2 instances in real time (install the CloudWatch agent in the EC2, can also install the agent on premise)

§ Monitor CloudTrail logged events

§ By default, logs are kept indefinitely and never expire

§ Archive log data

§ Log route53 DNS queries

o CloudWatch Logs insights - allows you to interactively search and analyse your log data in CloudWatch Logs using queries

o CloudWatch Vended Logs - are logs that are natively published on by AWS services on behalf of the customer. VPC Flow Flogs is the first Vended log type that will benifit from this tiered model.

· CloudWatch Agent

o Collects more logs and system level metrics from EC2 instances and your on premise servers

o Needs to be installed

· Authentication and Access Control

o Use IAM users or roles for authenticating who can access

o Use Dashboard Permissions, IAM identity-based policies, and service-linked roles for managing access control

o A permissions policy describes who has access to what

§ Identity-Based Policies

§ Resource-Based Policies

o There are no CloudWatch ARNS for you to use in an IAM policy. Use an *(asterisk) instead of the resource when writing a policy to control access to CloudWatch actions

· Pricing

· Limits