#office-hours (2020-01)
“Office Hours” are every Wednesday at 11:30 PST via Zoom. It’s open to everyone. Ask questions related to DevOps & Cloud and get answers! https://cloudposse.com/office-hours
Public “Office Hours” are held every Wednesday at 11:30 PST via Zoom. It’s open to everyone. Ask questions related to DevOps & Cloud and get answers!
https://cpco.io/slack-office-hours
Meeting password: sweetops
2020-01-08
Office hours is happening today!
let me know if you have any questions
I’ll be joining late… I have to sit in a call with Ingram Micro…. for my sins.
@btai
@btai has joined the channel
This project runs as a Kubernetes Job that aims at killing pods when it thinks the cluster is unbalanced. You can run it once or as a Cron Job to run it periodically.
New Zoom Recording from our Public “Office Hours” on 2020-01-08 is now available.
2020-01-10
I don’t remember who asked about ECS and terraform yesterday
But we have some news
Check out this PR: https://github.com/terraform-providers/terraform-provider-aws/pull/11506
This is great
Oh wow, what a coincidence! We just talked about that
2020-01-14
Something for this weeks office hours perhaps.
I am currently trying to move some of our IaC logic out of the IaC repo and provide a module / tool to our application developers to provosion the databases for them incl. deployment of secrets in side of k8s. (See: https://github.com/helm-notifier/Terraform-Infrastructure/blob/master/02-step2/postgresHelmNotifier.tf)
If you have Ideas or are doing this already I would love to hear about it!
Suggest to stick with official style conventions
2020-01-15
docs: from todays k8s office hours: https://discuss.kubernetes.io/t/office-hours-for-january-15th/9245/3?u=castrojo
Was the CloudFlare demo in today’s office hours?
it is planned
Still happening?
yes, join now
I’ll do it when you join
@Igor
I think I might have an outdate invite/zoom link
TY. I’m on
Thanks for that demo @Erik Osterman (Cloud Posse), that was awesome
Thanks for the demo Erik.
Here’s a brief comparison of CCPA and GDPR for those interested https://www.bakerlaw.com/webfiles/Privacy/2018/Articles/CCPA-GDPR-Chart.pdf
2020-01-22
Quick video I made of the IaC demo. Feel free to poke holes in it
New Zoom Recording from our Public “Office Hours” on 2020-01-22 is now available.
2020-01-29
- DevOps Interview Questions. What are your favorites?
- https://github.com/bregman-arie/devops-exercises/
- https://enterprisersproject.com/article/2020/1/devops-engineer-interview-questions-2020
- Jenkins in 2020, Pros/Cons?
@Pierre Humberdroz PROMCEPTION!
our way currently is currently prometheus per ns -> master prometheus per cluster -> dedicated monitoring cluster prometheus -> elastic metric exporter to our elasticsearch cluster and after 14 days -> backup to blobstorage with a tool to get the data restored of a given time range.
Side project
that’s rad!
nice!
that is rad. where’d the case come from?
Very nice interview with the Prometheus core team on: https://changelog.com/podcast/375
@roth.andy I have questions for you about your Raspberry Pi cluster.
shoot
Are you using Kubernetes or K3s?
I started the project before K3s supported multi-master, so currently it uses kubeadm
Have you been deploying using manifest files or helm?
I am using k3s with k3sup
The cluster doesn’t work yet, I got busy with other stuff, but it would definitely use helm
OK cool, that happened to me as well. Started it last year and I just got through rebuilding it with k3s. I had set it up with kubeadm before but wanted to use something lighter.
My plan is to get prometheus-operator Helm chart to get it up and running.
Then setup Rook using USB keys on each of the raspberry pi
I missed what system you’re talking about, is this a codefresh change? re: the PV changes
yes
@btai has more details
FROM node:12-slim
RUN apt-get -y update && apt-get -y install git && apt-get clean && rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/* /usr/share/man/?? /usr/share/man/??_*
ENV SOURCES /app
COPY . ${SOURCES}
WORKDIR ${SOURCES}
RUN chown -R node:node ${SOURCES} \
&& chmod 777 /tmp
USER node
RUN npm install --production --quiet
EXPOSE 5000
CMD ["npm", "start"]
I think this is what you don’t want to do for maximum layer caching
right
so a step before the COPY . would be to only copy over the packages.json and then download/process the deps.
Also consider using npm ci instead of npm install
I am not familiar with npm ci; what does that do?
nice tip!
interesting @roth.andy, will test out npm ci , always looking to improve build times
jk we already do this https://stackoverflow.com/questions/58482655/what-is-the-closest-to-npm-ci-in-yarn
@Pierre Humberdroz
Split into 2 projects
FROM node:12-slim
RUN apt-get -y update && apt-get -y install git && apt-get clean && rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/* /usr/share/man/?? /usr/share/man/??_*
ENV SOURCES /app
WORKDIR ${SOURCES}
RUN chown -R node:node ${SOURCES} \
&& chmod 777 /tmp
USER node
EXPOSE 5000
CMD ["npm", "start"]
FROM theBase:someVersion
COPY ./packages.json ${SOURCES}
RUN npm ci
COPY . ${SOURCES}
Also, consider using webpack and nginx or something instead of npm start
First time I hear to use nginx per container ? Why would I want to do that @roth.andy? It also goes against the philsophy of 1 process per container IMO.
In regards to webpack what is the reasoning for that?
nginx would be the one process, instead of node. Webpack can build a production deployment package with minified javascript
So you wouldn’t actually need to run npm install or anything, since webpack just creates one big index.js
this is an API and not a frontend project.
ah
no worries.
the main benefit of webpack or gulp in the backend world is you do not have to require everything on startup which makes startup time a lot faster. And the ability to use newer features
(or just use a multi-stage docker build, no?)
Yep, assuming you have caching set up. Since I don’t have caching it’s faster to actually make a base image
good point.
Anybody tried out yarn 2 yet?
New Zoom Recording from our Public “Office Hours” on 2020-01-29 is now available.