text stringlengths 0 1.49k |
|---|
**Erik St. Martin:** Alright, any other shoutouts before we close this thing out? Alright, I will take that as a no. So a big thank you for everybody for being on the show, especially Aaron - thank you for coming on. |
**Aaron Hnatiw:** You're welcome, thanks for having me. |
**Erik St. Martin:** This has been really great, and as Aaron said, he's available for anybody who has questions and wants to learn more about security and writing better code. We'll drop all of his contact details - home address, phone number... |
**Brian Ketelsen:** His home address, his social security number... \[laughter\] Yeah! |
**Erik St. Martin:** We get all of that before we put people on the show. \[laughter\] It keeps them in line. |
**Aaron Hnatiw:** I forgot to send you the CVV code, I'll get that to you after. \[laughter\] |
**Brian Ketelsen:** Thank you. |
**Erik St. Martin:** So huge thank you to all of the listeners right now... You can find us on [Twitter](https://twitter.com/GoTimeFM), GoTime.fm online, and if you wanna be on the show, have suggestions for guests or questions, hit us up at [ping](https://github.com/GoTimeFM/ping). With that, we'll see you next week. ... |
**Brian Ketelsen:** Bye! |
**Carlisia Thompson:** Bye, this was fun! |
**Aaron Hnatiw:** Bye, everyone! |
• Introduction to Cloud Foundry as an enterprise platform as a service |
• Background on Pivotal and Cloud Foundry contributors Jason Keene and Andrew Poydence |
• Discussion of Loggregator, a log system for Cloud Foundry written in Go |
• History of Cloud Foundry's transition from Ruby to Go |
• Advantages of using Go, including simplicity and maintainability |
• How Go enables distributed teams to contribute to the project easily |
• Challenges of developing Loggregator in Go when it was less widely adopted |
• Rewrite of Loggregator code due to need for new features and scalability |
• Adoption of gRPC for messaging and its benefits (security, cost savings) |
• Scalability issues with large deployments of Cloud Foundry and Loggregator |
• Creation of "diode" ring buffer concept to prioritize message delivery |
• Use of HTTP/2 multiplexing in gRPC for efficient stream handling |
• Pooling connections and load balancing to manage scale and efficiency |
• Loggregator's goal is to have an opinionated log structure |
• Loggregator uses protocol buffers for strict messaging and enables generic consumers to pull data without knowledge of Cloud Foundry specifics |
• The system aims to distance itself from being specific to Cloud Foundry, with a v2 API that distills core metrics and messages |
• Bosh Deployment is used to manage Cloud Foundry deployments, including on laptops for development purposes |
• The team discusses the project's use of Go and whether Generics would be useful in implementing certain data structures |
• Compiler limitations and workarounds in Go |
• Generics discussion, including solicitation of use cases by the Go team and potential impact on readability and maintainability |
• Code generation and working around generic type issues |
• Use of gRPC for messaging, including native implementation in Go, ease of use, and compatibility with protobufs |
• gRPC upgrade paths |
• Deprecation timelines for software components |
• Trade-offs between fast deployment and stability in production systems |
• DevOps movement and balance between ops and dev responsibilities |
• Operations knowledge for developers |
• Site Reliability Engineering (SRE) principles and practices |
• Cloud Foundry deployment scenarios and on-call experiences |
• Reliability metrics and measuring success in distributed systems |
• Setting SLA and SLI levels and understanding tradeoffs between reliability and other goals |
• Measuring message reliability and using an "error budget" in a streaming service environment |
• Project releases, including Contour for Envoy-based Ingress Controller |
• GoTTY project for sharing terminal sessions through web pages |
• G.E.R.T project for running Go on ARMv7 systems |
• GoScan tool for scanning IPv4 subnets and discovering hostnames |
• Discussing Go programming language tools and libraries |
• Upcoming conferences: dotGo, GopherCon Brazil, and Women Who Go in Paris |
• Speaking engagements at conferences (Carlisia Thompson and Brian Ketelsen) |
• NVIDIA's nvidia-docker project for container support on GPU hardware |
• #FreeSoftwareFriday shoutouts to present tool from the Go team and Concourse CI |
• Concourse CI: discussed as a tool for automating pipelines and tasks |
• eBPF (extended BPF): kernel technology for high-performance monitoring, mentioned with Go bindings through gobpf |
• dep: discussed as a vendoring tool for Go projects, praised by several participants |
• Legacy source code management techniques, such as using Git submodules to vendor dependencies |
• Skype call issues and humorous discussion about having successful co-guests |
**Erik St. Martin:** Welcome back, everybody, to another episode of GoTime. Today's episode is number 61. On the show today, your hosts are myself, Erik St. Martin, Carlisia Pinto is also here... |
**Carlisia Thompson:** Hello! |
**Erik St. Martin:** And Brian Ketelsen, who is half alive, half dead... |
**Brian Ketelsen:** Half dead. |
**Erik St. Martin:** \[laughs\] And our special guest for today -- we actually have two, both working on Cloud Foundry, is Jason Keene and Andrew Poydence. |
**Jason Keene:** Hey! |
**Andrew Poydence:** Hello! |
**Erik St. Martin:** So I guess maybe give a little bit of background... [Pivotal](https://pivotal.io/) and [Cloud Foundry](https://www.cloudfoundry.org/) kind of do a lot of stuff, so do you wanna maybe give a little bit of background about yourselves and the particular areas within Cloud Foundry that you are focused? |
**Jason Keene:** Yeah, so I joined Pivotal about two years ago now, and I work explicitly on Cloud Foundry. Cloud Foundry, for people who don't know, is an enterprise platform as a service, similar in style to Heroku, but it's on-prem. You can also run it in the cloud, in Google, or AWS, or whatever. So yeah, that's ki... |
**Andrew Poydence:** I joined over three years ago, and started on [Loggregator](https://github.com/cloudfoundry/loggregator), which is the log system for Cloud Foundry, where if you have an application running in Cloud Foundry, your `STDOUT` and `STDERR` could be gathered and shipped off to you, as the developer. That... |
**Erik St. Martin:** And just for like a separation of voices, that was Jason who spoke first, and Andrew second, right? |
**Jason Keene:** Yeah, yeah. |
**Brian Ketelsen:** So Loggregator is written in Go? |
**Jason Keene:** Yeah, so I would say a good majority of components in the Cloud Foundry system are implemented in Go now. Cloud Foundry has kind of a history of Ruby, but it went through multiple transitions where components were kind of incrementally re-written in Go... So our load balancer, reverse proxy component i... |
**Brian Ketelsen:** What lead to the moving from Ruby to Go? |
**Jason Keene:** I totally wasn't part of the company back then, but... Andrew, do you have any context there? |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.