Monitoring at the Speed of Cloud Native Why Kubernetes + Automated Monitoring Make a Winning Pair 

• Docker Captain • Docker Meetup Organizer • DevOpsDays Nashville Organizer • 14+ years in software development • 4+ years Docker experience Kevin Crawley Developer Evangelist Social Media @notsureifkevin (twitter, github, Instagram, linkedin, etc) Random Facts: • Rode a bicycle 100 miles in one day (ugh) • SciFi Nerd (current <3 is The Expanse) • I know how to play the Didgeridoo! 

Agenda • The “Official” State of DevOps • Measuring and Empowering Performance • Challenges of Monitoring Kubernetes + Cloud Native Applications • Considerations when monitoring container workloads • The Solution: Instana Automatic Monitoring • Instana Demo featuring Kubernetes Deployment • Q&A 

State of DevOps • High Performance vs Low Performance Organizations • How you implement cloud infrastructure matters • Software Delivery & Operational Performance (SDO) unlocks competitive advantages • Outsourcing by function hurts performance. • Key technical practices – like monitoring and observability – drive high performance 

Software Delivery Performance - Throughput Measuring the velocity of your application delivery organization 

Software Delivery Performance - Stability Measuring the resiliency of your application delivery organization 

Percentage of Work Done Manually ELITE PERFORMERS HIGH PERFORMERS LOW PERFORMERS Configuration Management 5% 10% 30% Testing 10% 20% 30% Deployments 5% 10% 30% Change approval process 10% 30% 40% https://puppet.com/resources/whitepaper/state-of-devops-report 

High Performance vs Low Performance Organizations High Performers • Deployments: > 1 hour and < 1 day • Lead Time for Changes: > 1 day and < 1 week • MTTR: < 1 day • Change Failure Rate: 0-15% Low Performers • Deployments: Once per week/month • Lead Time for Changes: > 1 month and <6 months • MTTR: > 1 week and < 1 month • Change Failure Rate: 46-60% https://puppet.com/resources/whitepaper/state-of-devops-report 

Challenges of Monitoring Kubernetes + Cloud Native Applications As more companies embark on digital transformation and cloud native technologies new and unique problems are surfacing: • Rapid introduction of performance problems and errors • Rapid introduction of new endpoints causing monitoring gaps • Lengthy root cause analysis as number of services expand 

Modern DevOps means eliminating manual work • More software is updated or added more frequently on more infrastructure • Good for business, difficult on conventional monitoring tools • Most monitoring tools employ many, if not all, of the following processes: • Manually write data collectors • Manually instrument code for tracing • Manually configure data collectors • Manually discover dependencies • Manually decide how to correlate data • Manually build dashboards to visualize correlation • Manually configure alerting rules and thresholds • Manually build data collection to store your metrics 

Speed Requires Automation • CI/CD automates your software delivery process • Manual steps slow you down! GitHub Puppet Labs Chef Selenium Kubernetes Jenkins 

Requirements for Automatic Monitoring • Zero or Minimal Configuration for the Automatic Discovery of Infrastructure and Software Components • Automatic instrumentation and tracing of every component in your application • Pre-existing alerts for supported technologies and frameworks • High resolution metrics and analytics to power Machine Learning Algorithms Automated continuous monitoring will keep your continuous deployment pipelines flowing smoothly and efficiently. 

Instana Automatic Monitoring Instana Agent: • One agent deployed once per host • Continuous automatic discovery of technology • Automatic metric collection • Automatic tracing • Automatic dependency mapping • Automatic support of over 200+ technologies, including Kubernetes, Mesos and OpenShift Continuous real time discovery and monitoring of ALL components Automatic No Plugins No Configuration 

DevOps Demands Automatic Don’t slow down Let the robot do the work! Automation Detect Capture Analyse (AI) Actionable Information Optimisations Troubleshooting Accelerate Delivery Incidents 

"Observability aims to provide highly granular insights into the behavior of systems along with rich context, perfect for debugging purposes." Cindy Sridharan Observability: The New Requirement 

Important considerations when monitoring container workloads • Monitoring distributed workloads isn’t optional • Distributed tracing isn’t optional either • The “true” cost of monitoring is only apparent when you don’t have it 

Monolithic vs Microservice courtesy of https://www.weave.works 

Monolith vs Microservice Monolith • Changes are Infrequent and Large • Impact of outage is widespread and pronounced • Monitoring requirements are fairly static Microservice • Small Changes are pushed frequently • Impact of outage is isolated • Requires dynamic and automatic monitoring solutions 

Monolith Outage 

Microservice Outage 

Deep Performance Analysis • Rise in Latency and Processing Time • DBO causing log(n) rise in latency and processing • Application Trace to Database led us to the offending endpoint • Fix deployed and improvement observed immediately (Next Slide) 

Immediate Feedback 

No content

No content

(and follow me on twitter @notsureifkevin) See how easy it is for yourself! Sign up for a Free Trial @ https://instana.com