Shashi Raina [AWS] & Al Sargent [InfluxData] | Build Modern Monitoring with InfluxDB and AWS | InfluxDays Virtual Experience London 2020

Shashi Raina
Partner Solution Architect, AWS
Al Sargent,
Sr. Director, Product Marketing,
InfluxData
Build modern
monitoring with
InfluxDB and AWS

Monitoring AWS: Best Practices
Monitoring AWS with InfluxDB
Configure Telegraf to monitor AWS
Demo

Best Practices for monitoring AWS

© 2020 InfluxData. All rights reserved. 4
So Why Monitor In the First Place?
To Gain Insights!
• Customer Experience
• Performance & Cost
• Trends
• Troubleshooting & Remediation
• Learning & Improvement

What Goes Into a Monitoring Plan?
Alerts
System
Knowledge
People
Actions
Tools

Alerting Best Practices
• Break alert crafting into batches. Highest Priority First
• Refine quickly.
• Alert to prompt an action
• Descriptive alerts to aid in prompt resolution
• Don’t only use email

Summary
- Check your monitoring approach
- Is it user-centric?
- Are you measuring the right things?
- Write a monitoring plan
- Start monitoring, test and iterate
The reason operations exists is to support the needs of the business.

© 2019 InfluxData. All rights reserved.11
Accumulate Act
Telegraf Inputs
CloudWatch plugin
ECS plugin
System plugin
Docker plugin
Kubernetes plugins
Kinesis plugin
MQTT & Modbus plugins
Flux
RDS joins
93 AWS services
AWS ECS &
Fargate
AWS EC2
AWS EKS
AWS Kinesis
IoT devices &
sensors
AWS RDS
MariaDB, MySQL,
Postgres
Analyze
AWS Global Infrastructure
InfluxDB Cloud
InfluxDB
Enterprise
InfluxDB
Purpose-built Time Series Database
Realtime Data Stream Processing
Visualization & Dashboarding
Data Analysis & Anomaly Detection
Alerting & Notifications
Alerting Systems
PagerDuty
Slack
Webhooks
Telegraf Outputs
AWS Kinesis
AWS CloudWatch
Grafana
Client Libraries
AWS Marketplace Billing

Telegraf input plugins on GitHub
Telegraf Inputs
CloudWatch plugin
ECS plugin
System plugin
Docker plugin
Kubernetes plugins
Kinesis plugin
github.com/influxdata/telegraf/tree/master/plugins/inputs

Setup telegraf.conf for CloudWatch

Accumulate Act
Telegraf Inputs
ECS plugin
System plugin
Docker plugin
Kubernetes plugins
Kinesis plugin
Flux
RDS joins
AWS ECS &
Fargate
AWS EC2
AWS EKS
AWS Kinesis
IoT devices &
sensors
AWS RDS
MariaDB, MySQL,
Postgres
Analyze
AWS Global Infrastructure
InfluxDB Cloud
InfluxDB
Enterprise
InfluxDB
Alerting Systems
PagerDuty
Slack
Webhooks
Telegraf Outputs
AWS Kinesis
AWS CloudWatch
Grafana
Client Libraries
CloudWatch plugin93 AWS services

First, the overall agent config
[agent]
# Run telegraf with debug log messages?
debug = false
# How often to collect metrics
interval = "30s"
# Default flushing interval for all outputs
flush_interval = "10s"
# How many metrics to cache
metric_buffer_limit = 50000

Specify your cloud region
[[inputs.cloudwatch]]
# Specify your AWS Region
# region = "eu-central-1" # Frankfurt
# region = "eu-north-1" # Stockholm
# region = "eu-west-1" # Dublin
# region = "eu-west-2" # London
# region = "eu-west-3" # Paris
# region = "eu-south-1" # Milan
# region = "us-east-1" # Virginia
# region = "us-east-2" # Ohio
# region = "us-west-1" # Northern California
region = "us-west-2" # Oregon
# https://docs.aws.amazon.com/AmazonRDS/latest/UserGuide/Concepts.RegionsAndAvailabilityZones.html

Specify your AWS credentials
# AWS credentials
# Credentials are loaded in the following order
# 1) Assumed credentials via STS if role_arn is specified
# 2) explicit credentials from 'access_key' and 'secret_key'
# 3) shared profile from 'profile'
# 4) environment variables
# 5) shared credentials file
# 6) EC2 Instance Profile
# access_key = ""
# secret_key = ""
# token = ""
# role_arn = ""
# profile = ""
shared_credential_file = "./credentials"
[default]
aws_access_key_id = AKIAI53FASDFP7J3KQ
aws_secret_access_key = 4EZ7As/Lmr2d1JgUaIdakr+58hpBJ
credentials

Specify your collection timing
# Requested CloudWatch aggregation Period (required - must be a multiple of 60s)
period = "1m"
# Collection Delay (required - must account for metrics availability via CloudWatch API)
delay = "5m"
# Recommended: use metric 'interval' that is a multiple of 'period' to avoid
# gaps or overlap in pulled data
interval = "1m"

Configure your metric namespaces
# Metric Statistic Namespace (required)
# List of namespaces:
https://docs.aws.amazon.com/AmazonCloudWatch/latest/monitoring/aws-services-cloudwatch-
metrics.html
namespace = "AWS/EC2"
# Maximum requests per second. Note that the global default AWS rate limit is
# 50 reqs/sec, so if you define multiple namespaces, these should add up to a
# maximum of 50.
ratelimit = 25

docs.aws.amazon.com/AmazonCloudWatch/latest/monitoring/aws-services-cloudwatch-metrics.html

Add tags to find your metrics more easily
# Optional tags that you can add
[inputs.cloudwatch.tags]
plugin = 'cloudwatch'
aws_service = 'ec2'

Specify the metrics you want Telegraf to pull
[[inputs.cloudwatch.metrics]]
# List of EC2 metrics available:
# https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/viewing_metrics_with_cloudwatch.html
names = ["StatusCheckFailed","EBSReadBytes","CPUSurplusCreditsCharged","EBSByteBalance%",
"StatusCheckFailed_System","EBSWriteBytes","NetworkIn","CPUCreditUsage",
"EBSIOBalance%","EBSReadOps","CPUCreditBalance","StatusCheckFailed_Instance",
"CPUUtilization","NetworkOut"]

List of metrics varies by AWS service

Specify your instance
[[inputs.cloudwatch.metrics.dimensions]]
name = "InstanceId"
# This will be unique for each AWS instance
value = "i-06025f2c26acfbf47"

Best practice: Have Telegraf monitor itself
# Collect metrics on Telegraf itself
[[inputs.internal]]
collect_memstats = true
# Tag stats with the metric name for easier retrieval
[inputs.internal.tags]
plugin = 'internal'

Send data to your InfluxDB instance
[[outputs.influxdb_v2]]
# Location of your InfluxDB Cloud instance
# Cloud URLs: https://v2.docs.influxdata.com/v2.0/reference/urls/
urls = ["https://us-west-2-1.aws.cloud2.influxdata.com"]
# Store token in an environment variable called US_WEST_2_1
token = "$US_WEST_2_1"
# Your org is the email you signed up with
organization = "asargent@influxdata.com"
bucket = "aws"
# About 5x faster
content_encoding = "gzip"

You can dual-write to multiple instances
[[outputs.influxdb_v2]]
# Cloud URLs: https://v2.docs.influxdata.com/v2.0/reference/urls/
urls = ["https://eu-central-1-1.aws.cloud2.influxdata.com"]
# Store token in an environment variable called EU_CENTRAL_1_1
token = "$EU_CENTRAL_1_1"
organization = "asargent+aws-eu-central-1@influxdata.com"
bucket = "aws"
content_encoding = "gzip"

To troubleshoot, write line protocol to stdout
[[outputs.file]]
files = ["stdout"]
data_format = "influx"

github.com/alsargent/telegraf-cloudwatch

Accumulate Act
Telegraf Inputs
CloudWatch plugin
ECS plugin
System plugin
Docker plugin
Kubernetes plugins
Kinesis plugin
Flux
RDS joins
93 AWS services
AWS ECS &
Fargate
AWS EC2
AWS EKS
AWS Kinesis
IoT devices &
sensors
AWS RDS
MariaDB, MySQL,
Postgres
Analyze
InfluxDB
Alerting Systems
PagerDuty
Slack
Webhooks
Telegraf Outputs
AWS Kinesis
AWS CloudWatch
Grafana
Client LibrariesAWS Global Infrastructure
InfluxDB Cloud
InfluxDB
Enterprise

InfluxDB on AWS

Shashi Raina [AWS] & Al Sargent [InfluxData] | Build Modern Monitoring with InfluxDB and AWS | InfluxDays Virtual Experience London 2020

More Related Content

What's hot (20)

Similar to Shashi Raina [AWS] & Al Sargent [InfluxData] | Build Modern Monitoring with InfluxDB and AWS | InfluxDays Virtual Experience London 2020 (20)

More from InfluxData (20)

Recently uploaded (20)

Shashi Raina [AWS] & Al Sargent [InfluxData] | Build Modern Monitoring with InfluxDB and AWS | InfluxDays Virtual Experience London 2020