Optimizing Your Container Environment: Monitor Pets and Cattle

Optimizing Your Container Environment: Monitor Your Pets and Cattle

Real-time Log Management and Analytics Built for the Cloud.

ServicesApplications

OperatingSystem

VirtualMachine

Hypervisor Machine

Containers are changing how software development and operations teams design, build and monitor systems. These containers - or “pets” and “cattle” - need consistent monitoring and optimization.

Containerization of environments regularly results in systems with large numbers of dynamic and ephemeral instances that autoscale to meet demands on system load.

GROWTH AND POPULARITY Docker Container Downloads

(in Millions)

December 2014 April 2015 June 2015

4xin six months

SECOND-MOST POPULAR OPEN SOURCE PROJECT

(Survey of more than 550 respondents conducted by Linux.com and The New Stack)

Fewer respondents said they contribute to Docker but it’s the project everyone said they want to

contribute to more.

0

50

100

150

200

250

300

350

400

KVMDockerOpenstack

Google's data centers and

applications are 100% container-

based — spinning up 2 billion

containers per week

The potential of container technology to unseat VMs threatens the market for products sold by Citrix, VMware, Dell, Hewlett-Packard, IBM and Microsoft.

$4B MARKET THREATENED

Amazon, Google, IBM, and Microsoft have all partnered with Docker and are building their own container-based solutions.

“Why is it such a big deal? Because it means genuine application

portability -- using lightweight packages instead of full VMs.“

Eric Knorr, InfoWorld

CONTAINER TYPES: PETS AND CATTLE

PetsContainers With Very Short Life Spans

• Require great attention

• Given actual names

• Unique and procured individually

• Sad if they die

• Common pets include MSSQLdatabases, SharePoint, legacy apps, and Unix systems

CONTAINER TYPES: PETS AND CATTLE

CattleContainers With Long Uptimes

• Look after themselves

• Don't have specific names - given numbers

• Managed in herds

• Can be simply replaced if they die

• Standardized

CONTAINER PROBLEMS + SOLUTIONSContainer Health: 6 Questions DevOps Need to Answer

How do you monitor your container’s resource utilization?

How do you get insight into the current and historical state of clusters?

How do you preempt container performance issues?

Do you know when a container dies or gets paused?

Are you able to see if a container went down in retrospect when troubleshooting?

How easy is it for you to capture CPU, network and memory utilization per container?

A specialized Docker container for log collection and monitoring

Centralized logging

capabilities for Docker

environments

Container-level resource usage

statistics such as CPU, Memory, Network, etc.

Monitoring Tools: Out-of-the-box saved searches, tags, (real-time) alerts and data visualizations

Preventive Tools:

Proactive alerts

CONTAINER CARE SOLVED: MONITORING

LOGGING ON DOCKER

The Logentries container is a dedicated logging container that runs alongside your existing Docker containers. Using Logentries you can collect and monitor container-level stats and logs from within each container and easily scale to collect logs from new containers without requiring additional configuration.

• Automatically collect, monitor and analyze container-level logs and stats

• Automatically scale your container environment without configuration

• Easily correlate container-level logs, stats, Docker API events, and host metrics

+

Sources: • InfoWorld, “9 key enterprise tech trends for 2015 and beyond” • WSJ, “Software Firms Scramble to Jump Into Containers” • James Thomason, “2015 Year of the Whale and other disruptive trends in IT” • SiliconAngle, “Open source: Docker’s secret sauce | #DockerCon” • Docker, “About Docker”

Additional Resources: • The State of Logging on Docker •

https://blog.logentries.com/2014/03/the-state-of-logging-on-docker/ • What is the Docker Stats API •

https://blog.logentries.com/2015/02/what-is-the-docker-stats-api/ • Docker Log Management Using Fluentd •

http://jasonwilder.com/blog/2014/03/17/docker-log-management-using-fluentd/

• How to Centralize Logs from CoreOS Clusters • https://blog.logentries.com/2015/03/how-to-centralize-logs-from-coreos-clusters/

Get started with container monitoring.@TrevParsons linkedin.com/in/trevparsons

Real-time Log Management and Analytics Built for the Cloud.