talk-data.com talk-data.com

Event

Kubernetes & Cloud Native Berlin Meetup February Edition

2024-02-29 – 2024-02-29 Meetup Visit website ↗

Activities tracked

2

Kubernetes & Cloud Native Berlin Meetup is happy to co-organize a meetup with the Mercedes-Benz.io team at Berlin Mitte!

---------------------------------------------

SCHEDULE:

[18:00] Doors Open

[18:20 - 18:30] Benazir Khan (Community Program Manager, Microsoft) Opening Remarks, Code of Conduct, and Cloud Native Rejekts coming up on March 17-18!

[18:30 - 19:00] Nele Uhlemann (Developer Advocate, Fiberplane) "Handling incidents collaboratively is like solving a Rubik's Cube"

[19:00 - 19:45] Break - time for networking with pizzas and beverages - thanks to our friends from Fiberplane for sponsoring our pizzas and Mercedes-Benz.io for beverages! :)

[19:45 - 20:15] Kai Lüke (Senior Software Engineer, Microsoft) "Deploying software and configuration with systemd-sysext and systemd-confext"

[20:15 - 21:30] More time to network with peers and colleagues from the industry at a relaxed pace

---------------------------------------------

TALK DETAILS:

"Handling incidents collaboratively is like solving a Rubik's Cube," Nele Uhlemann Abstract: Understanding the business outcome and the overall functionality of a system consisting of multiple distributed services and the infrastructure components to run them at scale is almost like solving a Rubik's Cube. Once an incident occurs, it is not enough to look at the single side of a Rubik's Cube. To solve the puzzle, all sides of the cube need to be considered. Similarly, when solving an incident, a collaboration of different teams is needed.

Administering and monitoring a distributed system should not be the single effort of a single engineering team. Observability should be a goal and have value for all engineering teams. Nevertheless, it is often a mantra just for SRE teams.

Coming from the perspective of an application engineer, I will outline how an application engineer benefits from understanding infrastructure and common incidents and how SRE teams can benefit from understanding common failures when talking about the application code. Let’s take a deeper look at what collaboration across different engineering teams means and how it supports the process of resolving the Rubik's Cube together.

---------------------------------------------

"Deploying software and configuration with systemd-sysext and systemd-confext," Kai Lüke Abstract: Using an image-based OS brings advantages and challenges. One challenge is the customization of a read-only image with additional host-level software and configuration, and how to manage this customization through the lifetime of a machine. It turns out that configuration management has similar problems to solve because reliable reconfiguration often struggles with atomic updating and cleaning of old state.

Instead of placing and managing a bunch of config files in /etc and binaries in /opt, we can now extend the filesystem hierarchy at runtime through overlay mounts of extension images with systemd-sysext and systemd-confext. Extension images can be supplied by the user, by the OS vendor, or third parties. They are either bound to the OS version or may be independently updated.

The use of single filesystem image files mounted at runtime allows to reliably manage the changes compared to unpacking files to the root filesystem and trying to keep track of them. For an immutable OS the runtime overlay also allows to bring in deeper changes that previously required a custom image build when changing contents on the read-only /usr partition. This presentation will share use cases for extension images in Flatcar Container Linux, as well as possible uses with general purpose distros and for configuration management. It will cover rough edges, conceptual limitations, and future improvements.

---------------------------------------------

We have limited seats, so RSVP at the earliest to secure a spot!

Sessions & talks

Showing 1–2 of 2 · Newest first

Search within this event →

"Deploying software and configuration with systemd-sysext and systemd-confext"

2024-02-29
talk
Kai Lüke (Microsoft)
Linux

Using an image-based OS brings advantages and challenges. One challenge is the customization of a read-only image with additional host-level software and configuration, and how to manage this customization through the lifetime of a machine. It turns out that configuration management has similar problems to solve because reliable reconfiguration often struggles with atomic updating and cleaning of old state.

Instead of placing and managing a bunch of config files in /etc and binaries in /opt, we can now extend the filesystem hierarchy at runtime through overlay mounts of extension images with systemd-sysext and systemd-confext. Extension images can be supplied by the user, by the OS vendor, or third parties. They are either bound to the OS version or may be independently updated.

The use of single filesystem image files mounted at runtime allows to reliably manage the changes compared to unpacking files to the root filesystem and trying to keep track of them. For an immutable OS the runtime overlay also allows to bring in deeper changes that previously required a custom image build when changing contents on the read-only /usr partition. This presentation will share use cases for extension images in Flatcar Container Linux, as well as possible uses with general purpose distros and for configuration management. It will cover rough edges, conceptual limitations, and future improvements.

"Handling incidents collaboratively is like solving a Rubik's Cube"

2024-02-29
talk

Understanding the business outcome and the overall functionality of a system consisting of multiple distributed services and the infrastructure components to run them at scale is almost like solving a Rubik's Cube. Once an incident occurs, it is not enough to look at the single side of a Rubik's Cube. To solve the puzzle, all sides of the cube need to be considered. Similarly, when solving an incident, a collaboration of different teams is needed.

Administering and monitoring a distributed system should not be the single effort of a single engineering team. Observability should be a goal and have value for all engineering teams. Nevertheless, it is often a mantra just for SRE teams.

Coming from the perspective of an application engineer, I will outline how an application engineer benefits from understanding infrastructure and common incidents and how SRE teams can benefit from understanding common failures when talking about the application code. Let’s take a deeper look at what collaboration across different engineering teams means and how it supports the process of resolving the Rubik's Cube together.