Advertisement · 728 × 90
#
Hashtag
#VelocityConf
Advertisement · 728 × 90

Ugh, speaker feedback for #VelocityConf is... useless because of low feedback counts, usually the most disgruntled people.

There are... a total of ~5 reviews across my workshop & talk, who hated them. I've spoken to or seen tweets from >>5 people who really enjoyed the talk.

0 0 1 0

Hey #agilecitybrs, I just got here from #velocityconf and have tons of insights to share from the SRE & DevOps worlds!

0 0 1 0
Post image

Great chatting with @joatmon08 as @joelaha & I wrapped up #velocityconf!

0 0 1 0
Post image Post image

Think about how you want to test your infra, and then try it! @joatmon08 #velocityconf http://joatmon08.github.io/#slides

0 0 0 0
Post image Post image Post image

“Does this actually work the way I think it should?” @joatmon08 on testing your infrastructure. #velocityconf

0 0 0 0
Post image Post image

“It’s just a JSON with a list of planned changes!” Accessible and informative demo (with clear explanations!) from @joatmon08 at #velocityconf.

0 0 0 0
Post image

Super cool that @joatmon08 is doing a @code liveshare for her #velocityconf talk demo (which will include @garethr’s conftest!) http://hashi.co/tdd-velocity

0 0 0 0
Post image Post image Post image

“Infrastructure testing is _hard_, and no one really talks about it.” @joatmon08 #velocityconf

0 0 0 0
Post image

Last talk of the day: @rynchantress introduces fellow “Hashicorporeal” @joatmon08. #velocityconf

0 0 0 0
Post image Post image

Platform choices for ING mean considering risk. @jannabrummel & Robin explain the benefits: chaos engineering, zero trust, and pipelines. #velocityconf

0 0 0 0
Post image

Detailed tool list for ING’s cloud migration: Azure, AKS, Terraform, @HelmPack, and more. Makes sense for them to use services, as a small team providing a platform to their wider org. @jannabrummel & Robin #velocityconf

0 0 0 0
Post image

“We’re not doing this [removing toil] only for our team; we try to do this for all teams we support.” Robin on running a platform for ING, at #velocityconf.

0 0 0 0
Post image

Great overview from @jannabrummel on why service mesh makes sense for ING. #velocityconf

0 0 0 0
Post image

Public cloud: so a bank doesn’t have to spend time on infrastructure development. @jannabrummel #velocityconf

0 0 0 0
Post image Post image Post image

“Rethink everything. Should I do this? Is it my core business?” @jannabrummel on cloud decisions at ING. #velocityconf

0 0 0 0
Preview
Honeycomb: Observability for Distributed Services Honeycomb is the only observability platform you need. Get all your data in one unified platform with limitless possibilities.

All my slides can be found at http://hny.co/liz and @cyen's slides can be found at http://hny.co/christine

Until next time! #velocityconf

0 0 0 0
Preview
Observing and understanding distributed systems with OpenTelemetry - O'Reilly Velocity Conference in Berlin 2019 Liz Fong-Jones and Yoshi Yamaguchi walk you through a microservice-based system with distributed tracing using OpenTelemetry. You'll progress from inspecting raw trace data to visualizing the d...

If you enjoyed my talks/workshops (and/or have feedback for me), please please please rate them. It's how I grow as a speaker :)

https://t.co/5PLFXM68s1 & https://t.co/dIKkPNrY0Y #velocityconf https://t.co/VZ0BuFNKmP

0 0 1 0
Post image

Interested in a last-minute @honeycombio demo, including of BubbleUp and our SLO feature? I'll be in the area near Cafe Cube with the dance music for the next hour. #velocityconf

0 0 0 0

They're working on building more safety rails. [ed: they're... reinventing Sentinel from scratch]

What's in their future? Prometheus, WeaveWorks Flux, and service meshes. #velocityconf

0 0 0 0
Post image Post image

In older versions of k8s, calls to create storage weren’t idempotent (which is confusing and expensive) - they are now, but node plugins aren’t. Any task using volumes can end up in a weird state. @endocrimes #velocityconf

0 0 0 0

They run the terraform commands inside of a CI/CD bash script they wrote themselves; it pulls the S3 state file, detects what changes to run, etc. [ed: aha, yup, homebrew.]

they do have a manual approval stage but would like to automate more. #velocityconf

0 0 1 0

Devs are encouraged to write their own terraform, with approval from infra, but most things are done through variable substitutions.

Terraform AWS token credentials are locked down to not be able to modify VPCs etc to prevent accidents #velocityconf

0 0 1 0

Supporting infra such as RDS databases: they use Terraform almost exclusively to provision _everything_ from users to monitoring to databases.

reuse as modules across multiple different regions and environments. CD platforms that auto-apply [ed: TFE? custom?] #velocityconf

0 0 1 0
Post image

Distributed heartbeat is what @endocrimes is talking about, but I also see a control plane and worker that care about each other.💖 #velocityconf

0 0 0 0

They provide a standard base Chart template, that all teams use. there are feature flags/conditional dependencies people can opt into. but some features such as Limits cannot be bypassed. #velocityconf

0 0 1 0

Delivery: Helm delivers "charts" which are packaged k8s manifests with versions and templating. #velocityconf

0 0 1 0
Post image

Retries won’t always work and your application may need to be rescheduled. [Also, I totally love the disclaimer.😂] @endocrimes #velocityconf

0 0 0 0

dex // federated openid connector by coreos, acts as a portal. they use Github as their IDP, but are moving to Okta in long term. those credentials are mapped to k8s credentials by Dex. #velocityconf

0 0 1 0
Post image Post image Post image

Using persistent volumes with containers: possible, but requires care! (Not enough storage vendors have APIs available via CSI yet.) @endocrimes #velocityconf

1 0 1 0

[ed: thank you for using inclusive language, @jmickey_!]

devs can set their own allowlists & blocklists, and routing rules, rather than needing someone to configure them for them. #velocityconf

0 0 1 0