DEVOPS WEEKLY ISSUE #418 - 30th December 2018
Happy new year to all the readers of Devops Weekly. See you all in 2019.
From our sponsor, Victorops
The VictorOps team hopes you’ve had a warm and wonderful holiday season and a happy new year! See how VictorOps on-call incident management software continues to help DevOps teams build and maintain reliable services in 2019:
http://try.victorops.com/devopsweekly/why-victorops
News
A detailed article on the perils of dashboards, and management and decision making only based on the data we can easily observe. Very relevant to the management of complex systems.
https://www.forbes.com/sites/davidshaywitz/2018/12/24/we-are-not-a-dashboard-contesting-the-tyranny-of-metrics-measurement-and-managerialism/#1ed157da315b
We’re increasingly aware of the complexity of the distributed systems we’re managing it feels. This post highlights the importance of metrics and dashboards when it comes to building an understanding of how your system actually runs.
https://utcc.utoronto.ca/~cks/space/blog/sysadmin/MetricsAndUnderstandingMore
An interesting analysis of the recent Kubernetes vulnerability. It focuses on how the same general issue can arise in many microservices applications, and how to avoid it by moving away from access control lists towards capability-based security models.
https://www.cloudatomiclab.com/confused-deputies/
Hypervisors are pretty cool, providing a different level of isolation from containers. This post posits that, rather than many clusters, the future of Kubernetes will be based on light-weight virtual machines.
https://tech.paulcz.net/blog/future-of-kubernetes-is-virtual-machines/
A useful deep dive into managing organizations in AWS. Discusses Service Control Poclies and how they can be used to control various services for the whole org.
https://www.awsadvent.com/2018/12/13/i-gots-to-get-organizized/
A look at moving from ElasticBeanstalk to Kubernetes. The driving factor is interesting - resource utilisation in pre-production environments and developer productivity.
https://medium.com/oyotech/from-elastic-beanstalk-to-kubernetes-stage-1-1ada720e301f
An interesting analysis of the relationship between developers and operators, focusing on the thorny issue of accountability. Based on a recent survey of 2000 practitioners.
http://resources.overops.com/wp-content/uploads/2018/12/Dev-vs.-Ops_The-State-of-Accountability.pdf.pdf
A series of patterns and anti-patterns around creating feedback loops in order to improve operational practices and processes. Good tips on meetings, training and more.
http://blog.beyond20.com/how-to-create-devops-feedback-loops
Events
Devopsdays New York is coming up on January 24th and 25th, with a program that covers shifting left security, value stream mapping, offshore teams, documentation, ethics and the usual open spaces sessions. Tickets are on sale now.
https://www.devopsdays.org/events/2019-new-york-city/welcome/
Tools
Goldpinger is a service for helping you understand and visualize the network in your Kubernetes cluster. It also exposes Prometheus metrics that can be alerted on.
https://github.com/bloomberg/goldpinger
ConPan provides a handy tool which analyzes debian-based Docker images, lists the install packages and then provides information about issues and known vulnerabilities.
https://github.com/neglectos/ConPan
The VictorOps team hopes you’ve had a warm and wonderful holiday season and a happy new year! See how VictorOps on-call incident management software continues to help DevOps teams build and maintain reliable services in 2019:
http://try.victorops.com/devopsweekly/why-victorops