#office-hours (2021-07)
“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
2021-07-07
@here office hours is starting in 30 minutes! Remember to post your questions here.
Erik Osterman (Cloud Posse) has joined Public “Office Hours”
Eric Berg has joined Public “Office Hours”
Vlad Ionescu has joined Public “Office Hours”
Thayne Trevenen has joined Public “Office Hours”
Frank Scalzo has joined Public “Office Hours”
Ossie Botu has joined Public “Office Hours”
Erik Osterman (Cloud Posse) has joined Public “Office Hours”
David Hawthorne has joined Public “Office Hours”
Ian Bartholomew has joined Public “Office Hours”
Ossie Botu has joined Public “Office Hours”
Florain Drescher has joined Public “Office Hours”
Antarr Byrd has joined Public “Office Hours”
Luis Masaya has joined Public “Office Hours”
Evan Pitstick has joined Public “Office Hours”
Yusuf Adeyemo has joined Public “Office Hours”
Jeremy (Cloud Posse) has joined Public “Office Hours”
vikram yerneni has joined Public “Office Hours”
Oliver Schoenborn has joined Public “Office Hours”
PePe Amengual has joined Public “Office Hours”
Erik Osterman (Cloud Posse) has joined Public “Office Hours”
Sorry everyone - my connection is too bad we had to end early.
No worries! It happens to everybody!
no issue, although I’m glad because I missed it
Thanks guys
2021-07-08
I’d like to talk about approaches to distinguishing between (datadog) metrics from our internal services (EKS) and the infrastructure services, like DD agent, k8-cni, etc.
Obviously, one approach is to add a tag to each resource that identifies it as internal. Interested in hearing others’ experiences
2021-07-09
I believe we’ve talked about this before on office hours but I’m going to bring it up again as it’s still a topic that I feel doesn’t get enough attention:
How to manage Terraform dev/stage/prod releases when you’re utilizing terraform workspaces > the directory structure approach?
For my largest terraform project (couple dozen root modules), we utilize develop , master , and release branches to add some release process around promoting changes. This works well in some regards, but it’s also difficult in that upgrading our automation to tf v1.0 for example requires that we roll everything forward as fast as possible because that change isn’t held in isolation on the branch.
Anyway, looking to refresh on this topic. And particularly hear about how CloudPosse handles it.
My company was using branches the same way when I arrived and I got us off it as fast as I could for those same reasons and moved to directory separation. It was an exhausting process managing promotion of changes across branches.
the most asked question, I would move away from branches and use folder separation, we also used dev,qa, and prod branches before, it was a nightmare. suggest to only use features short-life branches and PRs for review the code added. everyone is talking about mono-repo mono-branch as single source of truth, with daily PRs.
now the magic happen in your CICD, it should be smart to know which folder to apply in which order, you can control releasing by promoting changes through environments, with simple testing at the end of the delivery to make a perfect deployment.
I like the idea to apply git-flow into Terraform, but I guess only small part of it could be helpful in IaC
it was a nightmare
heavy emphasis on this
Yeah interesting that both of you have this position. I hear ya… though I don’t exactly know how I would make that fit into the SweetOps methodology. Will be a good topic for discussion.
2021-07-10
2021-07-12
question for next office-hours meeting: https://sweetops.slack.com/archives/CC38YGGUE/p1626093139033800
which was built as a reverse proxy
2021-07-14
@here office hours is starting in 30 minutes! Remember to post your questions here.
I’m curious what the test workflow looks like in the cloudposse Terraform repositories. Rather selfishly I want to finish up this PR and the feedback loop + testing methodology is keeping me from completing it (and thus I’m using a fork in the mean time). Not sure if this is the place to go about it.
Erik Osterman (Cloud Posse) has joined Public “Office Hours”
Vicken Simonian has joined Public “Office Hours”
Ossie Botu has joined Public “Office Hours”
Emile Fugulin has joined Public “Office Hours”
David Scott has joined Public “Office Hours”
Yuri Lima has joined Public “Office Hours”
Andy Miguel (Cloud Posse) has joined Public “Office Hours”
Sam C has joined Public “Office Hours”
Jailson Silva has joined Public “Office Hours”
Yusuf Adeyemo has joined Public “Office Hours”
Anere Faithful has joined Public “Office Hours”
Michael Holt has joined Public “Office Hours”
Matt Calhoun has joined Public “Office Hours”
Mohammed Yahya has joined Public “Office Hours”
Thayne Trevenen has joined Public “Office Hours”
Matt Gowie has joined Public “Office Hours”
Neil Gealy has joined Public “Office Hours”
imran.hussain has joined Public “Office Hours”
Brad Janke has joined Public “Office Hours”
Denys has joined Public “Office Hours”
Andy Roth has joined Public “Office Hours”
Heath Snow has joined Public “Office Hours”
Neil Gealy has joined Public “Office Hours”
Marc Slayton has joined Public “Office Hours”
Felipe Sakatauskas has joined Public “Office Hours”
Jailson Silva has joined Public “Office Hours”
Stevan Arychuk has joined Public “Office Hours”
shreenu kumar has joined Public “Office Hours”
Graziele Vasconcelos has joined Public “Office Hours”
Charles Sperbeck has joined Public “Office Hours”
Soham Dutta has joined Public “Office Hours”
Fernando Sanz has joined Public “Office Hours”
Tim Gourley has joined Public “Office Hours”
Has anyone heard any updates on when Kubernetes will fix the issue with Jobs and InitContainers?
Eric Berg has joined Public “Office Hours”
Nick James has joined Public “Office Hours”
Mohammed Yahya has joined Public “Office Hours”
Denys has joined Public “Office Hours”
emem umoh has joined Public “Office Hours”
Michael Jenkins has joined Public “Office Hours”
Blaise Pabon has joined Public “Office Hours”
Othman Musleh has joined Public “Office Hours”
Neil Gealy has joined Public “Office Hours”
PePe Amengual has joined Public “Office Hours”
If I remember correctly it had trouble with path based, but host based works great
thanks, yes I’m looking for path based
links from today’s session:
• https://aws.amazon.com/blogs/containers/aws-controllers-for-kubernetes-ack/
• https://www.bschaatsbergen.com/behind-the-scenes-lambda
• https://github.com/helm/community/blob/main/hips/hip-0011.md
• https://aws.amazon.com/blogs/security/how-to-relate-iam-role-activity-to-corporate-identity/
• https://github.com/hashicorp/terraform-provider-aws/issues/20137
• https://www.aquasec.com/news/aqua-security-acquires-tfsec/
Rizky Ramadhan has joined Public “Office Hours”
A V has joined Public “Office Hours”
| [Managing Customer KMS Keys | Cloud Posse Explains](https://www.youtube.com/watch?v=t3Ecgx8NYkM) |
RE: aws-controllers-k8s - IAM is not supported, this was the biggest bummer for us and we gave up, moved all to terraform
also if you read the thread - there’s really no way to implement it securely. You will end up giving a controller iam:* which is a huge hole
I think this is the main showstopper there
wow, pretty useless without the ability to create IAM. In the end, we’re back to using something like terraform.
2021-07-15
2021-07-19
2021-07-21
Hi all, new here. Looking forward to my first live office-hour talk. Cheers.
@here friendly reminder we have @Taylor Dolezal from HashiCorp demoing Waypoint today in office hours!
Hope to see you there
@here office hours is starting in 30 minutes! Remember to post your questions here.
How are people in our community handling drift In the Real World? Did anyone try driftctl? What “homebrew” solutions have community members in use at the moment?
Also found out about clairvoyance the other day but I haven’t personally been able to test run it myself.
https://github.com/reulan/clairvoyance
Via corresponding hashicorp talk here: https://www.youtube.com/watch?v=zlwhw3YGlUc
@oskar my team is using spacelift which has native drift detection. there is also an interesting connection between increasing drift and lack of environmental progression automation; in other words: the less automation between deployments from environment-to-environment…the more drift!
thanks for chipping in jonas i’ve heard good things about spacelift, thanks for recommending it. very much agree on the automation side - yet some contexts / apis / providers can at least run into some “transition time” and in some cases do simply provide better ux through their respective native gui interface in some maybe (?) rarer cases (specific example would be setting up an alert in newrelic or datadog). yet again i agree - eg a “gitops” cicd workflow definitely should best practice and a default (i.e. atlantis). by that there should - as a strong guideline - very seldom be any outside-of-terraform manipulation of infrastructure.
i’m just asking to be looking at it from this “exceptional” perspective. maybe the alert example would be a good one to talk about!
@oskar you lost me lol. what now?
tl;dr there (unfortunately) are scenarios where even over a longer time we can regularly expect drift - e.g. setting something up through a very user friendly web interface for a specific resource comes to mind (e.g. complex alert plus rule in datadog). how do people in the community handle those kind of cases?
that is literally what spacelift does.
well that’s one feature.
that’s awesome. will definitely look further into it for sure
right now we’re working on altering when drift_time > x.
ok what does that mean exactly? are you talking about “auto-syncing” things after a while?
(or semi-auto)
it is not configuration management or eventually consistent: no.
i see, will look into it thanks for the rec again
Erik Osterman (Cloud Posse) has joined Public “Office Hours”
Rupinder Dhariwal has joined Public “Office Hours”
Taylor Dolezal has joined Public “Office Hours”
Adam Blackwell has joined Public “Office Hours”
Scott Mathson has joined Public “Office Hours”
Ian Bartholomew has joined Public “Office Hours”
David Scott has joined Public “Office Hours”
Andy Miguel (Cloud Posse) has joined Public “Office Hours”
Jim Park has joined Public “Office Hours”
Jonas Steinberg has joined Public “Office Hours”
Patrick Joyce has joined Public “Office Hours”
Michael Jenkins has joined Public “Office Hours”
Mikael Fridh has joined Public “Office Hours”
Benjamin Smith has joined Public “Office Hours”
Ray Myers has joined Public “Office Hours”
Joel Castillo has joined Public “Office Hours”
Miles Monteleone has joined Public “Office Hours”
Thayne Trevenen has joined Public “Office Hours”
Antarr Byrd has joined Public “Office Hours”
Michael Genrich has joined Public “Office Hours”
Matt Calhoun has joined Public “Office Hours”
Andrew Thompson has joined Public “Office Hours”
17866946419 has joined Public “Office Hours”
Andrew Thompson has joined Public “Office Hours”
Oskar Maria Grande has joined Public “Office Hours”
David B has joined Public “Office Hours”
Luis Masaya has joined Public “Office Hours”
PePe Amengual has joined Public “Office Hours”
Jim Antoniou has joined Public “Office Hours”
Mazin Ahmed has joined Public “Office Hours”
Murali Krishna Koppuravuri has joined Public “Office Hours”
Felipe Sakatauskas has joined Public “Office Hours”
Satish U has joined Public “Office Hours”
Question for today’s office hours: How do you use Git pre-hooks for identifying secrets on organization-level? I’m looking for ideas to detect sensitive commits before it’s committed and pushed to Github. Any ideas how would you approach this on org level?
Winson Chan has joined Public “Office Hours”
Neil Gealy has joined Public “Office Hours”
Arjun Venkatesh has joined Public “Office Hours”
Julian Severino has joined Public “Office Hours”
James Haughey has joined Public “Office Hours”
Murali Krishna Koppuravuri has joined Public “Office Hours”
Scott Mathson has joined Public “Office Hours”
Adeoye R has joined Public “Office Hours”
Blaise Pabon has joined Public “Office Hours”
@David regarding your yaml question from earlier: I doubt waypoint natively reads YAML, but coincidentally that is precisely what cloud posse’s Atmos does. I assume you know that, but if not check it. It’s awesome.
So if you had the time you could get a yaml –> atmos(terraform) –> waypoint workflow going on. pretty sure.
@Andy Miguel (Cloud Posse) power outage
ack
I won’t return probably
Cc @matt
Erik Osterman (Cloud Posse) has joined Public “Office Hours”
Zachary Loeber has joined Public “Office Hours”
Yusuf Adeyemo has joined Public “Office Hours”
Jim Park has joined Public “Office Hours”
@Erik Osterman (Cloud Posse): waypoint looks like a serverless deployment tool (like OpenFAAS)… am I right?
From what I understood, waypoint is like serverless, but 10x more dynamic, can be broader for K8S and other deployments
heh, well, this is the enigma of waypoint, it’s different things to different people
IMO, it’s honestly flat-out a CI/CD platform. It’s self hosted. It offers some “providers” (e.g. jenkins plugins). It handles build. It handles deploy. It handles release.
So in otherwords, waypoint would be used to build, deploy, and release serverless apps, k8s apps, etc.
it presents a clean HCL DSL for defining it (as opposed to the YAML approach favored by circle, github actions, gitlab ci, etc)
It’s really interesting!! I’m definitely giving it a try, hopefully will be able to use it for production
Hey there, just chiming in to react to last week episode. Terragrunt knows how to provision S3 (and gcs) buckets by itself before triggering terraform. I love these podcasts, don’t stop!
Yes, but at the same rate, it’s kind of awkward that we’re using terraform for IaC and yet the statebucket is not controlled by terraform
modifying that state bucket (e.g for compliance) in Terragrunt requires a PR to terragrunt, vs managing it in terraform https://github.com/cloudposse/terraform-aws-tfstate-backend
2021-07-26
@here August 18th we will have another special edition of Office Hours!
@Taylor Dolezal will be returning to demo HashiCorp’s Boundary project. Please queue up your questions here and hope to see you there
| [Waypoint Demo (AWS Lambda) | Cloud Posse Guest Speaker: Taylor Dolezal](https://www.youtube.com/watch?v=Kl2d2JMtnOo) |
2021-07-27
| [Waypoint Demo 2 of 3 (Kubernetes) | Cloud Posse Guest Speaker: Taylor Dolezal](https://www.youtube.com/watch?v=8z852c0wtHY) |
Hey Y’all, I am going to start putting TF state files in S3, whats the best policy to be as safe as possible because TF hasn’t given us anything to help with secrets in FT state?
Don’t put secrets in TF :)
More specifically, move to using other mechanisms for secrets, such as AWS’s Secrets Manager.
| [Waypoint Demo 3 of 3 (Minecraft on EKS!) | Cloud Posse Guest Speaker: Taylor Dolezal](https://www.youtube.com/watch?v=GQrWmGvGOP4) |
2021-07-28
@here office hours is starting in 30 minutes! Remember to post your questions here.
no questions here today so far so i’m asking something a bit off-topic maybe: has anyone test-run opstrace as an alternative to other saas “o11y” offerings? if so how did it go and did you go further into production with it? thanks!
addendum for people who haven’t read about it yet. it’s very much centred around k8s - relying on cortex and loki. blog articles have a strong docker ~2013-ish vibe so beware - could be too good (marketing) for their own sake (yet).
https://github.com/opstrace/opstrace https://opstrace.com/blog/cloud-provider-integrations https://opstrace.com/blog/a-giant-leap-for-alerts
this looks like a whole grafana.com backend stack assembled in one project
yeah it would be awesome if only for helping run and maintain prometheus/grafana stacks but apparently can be much more and manageable (at least that seems to be the promise). pulling in cloud level events etc.
we’ve been doing cortex in-house for 1 year. A ton of efforts but it pays off. Grafana pricing for us was 10k+ month, while we packed it in under 1k
I’d love to see if that thing really gives a working opinionated setup
1
same same. very excited to learn about this project (again don’t know if it really is as good as it looks at first sight!)
Erik Osterman (Cloud Posse) has joined Public “Office Hours”
Thayne Trevenen has joined Public “Office Hours”
Brandon vh has joined Public “Office Hours”
Andy Miguel (Cloud Posse) has joined Public “Office Hours”
Oskar Maria Grande has joined Public “Office Hours”
David Hawthorne has joined Public “Office Hours”
Michael Jenkins has joined Public “Office Hours”
Tim Gourley has joined Public “Office Hours”
Andrew Thompson has joined Public “Office Hours”
Jonas Steinberg has joined Public “Office Hours”
Blaise Pabon has joined Public “Office Hours”
Andy Roth has joined Public “Office Hours”
Nick James has joined Public “Office Hours”
Clayton Olley has joined Public “Office Hours”
links from today’s session:
• https://www.hashicorp.com/blog/mitchell-s-new-role-at-hashicorp
• https://github.com/Cigna/confectionery
• https://github.com/jckuester/awsls
• https://news.ycombinator.com/item?id=27974077
• https://github.com/cloudposse/terraform-aws-tfstate-backend
Yusuf Adeyemo has joined Public “Office Hours”
Mykola Lev has joined Public “Office Hours”
Mykola Lev has joined Public “Office Hours”
PePe Amengual has joined Public “Office Hours”
Fernando Castillo has joined Public “Office Hours”
Luis Masaya has joined Public “Office Hours”
Fernando Castillo has joined Public “Office Hours”
Winson Chan has joined Public “Office Hours”
Marc Slayton has joined Public “Office Hours”
Joel Caceres has joined Public “Office Hours”
reminds me the most: https://www.kubernetic.com/
used for a long time - positive feedback
Andrew Thompson has joined Public “Office Hours”
This week in Jonas’ world
probably not as extensible as this open source chatbot framework though - have not put much time into it myself yet and i’m biased as a clojure fan. still: https://github.com/yetibot/yetibot
wow… yeti looks amazing
yeah it totally is. if you are only a little into lisp you should check it out. and if you are into babashka (https://github.com/babashka/babashka) definitely check it out!
I was raised on Lisp …but that was a very long time ago.
haha it’s like riding the bicycle - never forget it really
there is not much to forget actually ;D
Hao Wang has joined Public “Office Hours”
Sheldon Hull has joined Public “Office Hours”
Mosh has joined Public “Office Hours”
Question: would like to know if anyone has used PGP to encrypt AWS access keys generated by the TF module for secret keys… what’s your experience been like?
related to secrets discussion
uwaila adams has joined Public “Office Hours”
I wouldn’t be much of a fanboy if I didn’t bring up K9s when there’s talk about K8s management GUIs
David Hawthorne has joined Public “Office Hours”
question: how are people terraforming alerts specifically in the case where there are many identical underlying resources spread across numerous accounts or environments that require different threshold values? So to articulate a bit more: imagine you have 20 EKS clusters and you want to monitor something like host IO (or whatever): something that is going to vary from environment-to-environment and that you can’t just set defaults for. How are people handling this case? My approach is as follows, but I’d love to learn about something better or alternatives in general?
1. have a directory with yaml files each containing a single "alert"
2. in each yaml alert file have some non-environment specific values like "query" which will always be the same from cluster-to-cluster
3. have an "options" section which, basically/unfortunately, contains all the environmental specification, mainly the threshold values *for every environment*
4. loop through all these files using terraform fileset
5. yamldecode every file
6. put the name of the alert in a local tf object
7. add the non-environment specific options as a list
8. get all the environment specific stuff and add it to that object
9. iterate over all that with terraform each spinning up alerts in the process with names and environmental config
# example
name: istio-host-latency-alert
type: metric alert
query: |
avg(last_5m):top(top(avg:istio.mesh.request.duration.milliseconds.sum{cluster_name:${stage}} by {host}, 10, 'mean', 'desc')/top(avg:istio.mesh.request.duration.milliseconds.count{cluster_name:${eks_cluster}} by {host}, 10, 'mean', 'desc'),10,'mean','desc') > ${critical_threshold}
message: |
({{event.tags.cluster_name}}) Detected Host Latency Greater than 1 Second
escalation_message: ""
tags: []
options:
- dev:
notify_no_data: false
renotify_interval: 60
notify_audit: false
timeout_h: 60
include_tags: true
require_full_window: true
threshold:
critical: 1000
warning: 50
- qa:
notify_no_data: false
renotify_interval: 60
notify_audit: false
timeout_h: 60
include_tags: true
require_full_window: true
threshold:
critical: 1000
warning: 50
- client-test:
notify_no_data: false
renotify_interval: 60
notify_audit: false
timeout_h: 60
include_tags: true
require_full_window: true
threshold:
critical: 1000
warning: 50
...
looks sane to me. anything specific that bugs you there? if you look for other perspectives, i’m in the same mindspace unfortunately.
very good question thanks for sharing. i really also like the “dynamic” view @marc slayton has on this as well. and to paraphrase maybe what you got at for the “static” approach you painted here jonas - not too many layers with the parametrization as @Erik Osterman (Cloud Posse) put it. and maybe also disregarding some of those differences to even simplify further as @roth.andy has said. hope i got this right. either way, thanks again very interesting topic to me as well.
sound like good rules of thumb for the “static” view on this kind of “o11y” (charity majors probably disagrees indeed btw ).
ha, yeah.
sorry to hijack this thread again but @marc slayton somebody was asking about which APM solution you are using before the zoom closed. i’m also curious
Any chance to have a global multiplier for an environment? Maybe a set of multipliers, say one for timeouts, another for RPMs. They you’d customize only multipliers. Of course you still should have a bypass to override a single value, when needed.
I missed the office hours yesterday so didn’t see this until I was watching the recording but this is actually exactly the rabbit hole I started down just yesterday for AWS Cloudwatch Alarms. I’m approaching this using the CloudPosse YAML Config module and the built in Parameters variable.
The only thing I haven’t figured a solution out for yet is setting default parameter values if not set in var.parameters
This is what my YAML File looks like:
ConsumedReadCapacityUnits:
metric_name: "ConsumedReadCapacityUnits"
metric_namespace: "AWS/DynamoDB"
treat_missing_data: "ignore"
comparison_operator: ${ConsumedReadCapacityUnits_comparison_operator}
description: ${ConsumedReadCapacityUnits_description}
metric_value: ${ConsumedReadCapacityUnits_metric_value}
evaluation_periods: ${ConsumedReadCapacityUnits_evaluation_periods}
period: ${ConsumedReadCapacityUnits_period}
statistic: ${ConsumedReadCapacityUnits_statistic}
threshold: ${ConsumedReadCapacityUnits_threshold}
dimensions:
TableName: ${ConsumedReadCapacityUnits_TableName}
And I pass this into the YAML Config module:
parameters = {
ConsumedReadCapacityUnits_TableName = "terraform-registry-touching-gorilla"
ConsumedReadCapacityUnits_threshold = "1"
ConsumedReadCapacityUnits_period = "300"
ConsumedReadCapacityUnits_evaluation_periods = "1"
ConsumedReadCapacityUnits_metric_value = "1"
ConsumedReadCapacityUnits_description = "Alarms when ."
ConsumedReadCapacityUnits_statistic = "Average"
ConsumedReadCapacityUnits_comparison_operator = "GreaterThanOrEqualToThreshold"
}
Florain Drescher has joined Public “Office Hours”
@Erik Osterman (Cloud Posse) this is what I have automatically initialized on each repo. I also have the CI checks run this too. It’s not perfect but it helps.
Gitleaks with lefthook (I’ve stopped using pre-commit framework as lefthook is super fast). https://github.com/sheldonhull/ci-configuration-files/blob/06f539315716d7a04fbf4ffbfd2e53e906729ef1/lefthook.secrets.yml#L4-L7
Vicken Simonian has joined Public “Office Hours”
Git secrets is good too! It’s another layer and can be a global hook. I think “shift left” by having it checked locally is ideal. However, the final source is really the CI action so no one can bypass by accident
All my projects have a go-task/make command to start work. task init. This also installs the hooks
Not perfect, but this ensures any new repo has the required policies “self setup”
Mohammed Yahya has joined Public “Office Hours”
This is what my init command looks like. I never leave projects “barebones” I use a way less pretty version like build harness that ensures tools are setup with no fuss/complexity.
init:dev:
desc: initialize tools for a developer, but not required for CI
cmds:
- |
dotnet --list-sdks || echo -e "{{ .red}} :small_red_triangle: dotnet-tools not available. Please install manually the first time here: <https://dotnet.microsoft.com/download> and then run command again {{.nocolor}}"
mkdir -p {{ .TOOLS_DIRECTORY }} || echo -e "{{ .dark_gray}}:arrows_counterclockwise: skipped creating {{ .TOOLS_DIRECTORY }} directory per already exists {{.nocolor}}"
# Uninstall manually
export PATH="$PATH:{{ .HOME }}/.dotnet/tools"
dotnet tool install --global GitVersion.Tool || dotnet tool update --global GitVersion.Tool # && echo -e "{{.green}} :white_check_mark: gitversion tool installed{{.nocolor}}"
# If gitversion gives problems with docker commands then evaluate just running as dotnettool
lefthook install
see last line. The init = always sets up whatever i know is important
That’s from https://taskfile.dev/#/ which I use wherever I’d use make (cross platform, parallelism, file watcher, very simple structure, and basic templating). This is nice for basics. It’s my current goto though I’m exploring other things like Atmos too
if you remove the cross-platform requirement you seem to be having would you be able to unbiasedly recommend it over make still? even if you / your team knows make well?
thanks for sharing btw. i didn’t know taskfile.dev
Personally I think Make is the default simply because it’s always been there. I don’t see any advantage in it. Wasn’t intended as a devops task runner tool, was meant for C development, so lots of clunky work arounds for devops usage.
Imo, a single curl bootstrap of task means it’s super easy to get going and if I had to pick it would be a no brainer for me
got it, so to paraphrase to you a more modern take. i would agree.
You get basic templating too and cross platform support can’t be discounted even if no one uses it yet, cause you never know!
version: '3'
includes:
build: ./Taskfile_{{OS}}.yml
You can use templating and fingerprint work for incremental builds. https://taskfile.dev/#/usage?id=by-fingerprinting-locally-generated-files-and-their-sources
The CI jobs in 3 seoncs bootstrap with snap install –classic task or whatever it is and you are good.
that is nice. thanks, will check it out.
Give this a shot as a starter sometime.
It’s NOT perfect. If you have Python devs then use a python task runner, Go use mage, etc.. If you need something instead of Make though, it’s fantastic.
I did a starter write up (have more in the ci-configuration repo i linked in main room), but as I started I tried to note some of the nice base configurations I setup. I plan on improving too with vars.yml instead of embedding the color formatting into the main taskfile.
very nice - coming up on my reading list - thanks for sharing again
@Erik Osterman (Cloud Posse) If we get any time at the end of call, would love to know…
- Any starter repo for using variant to run workflow of terraform stacks?
Doing a quick day of work to try and see if I can get away from Terragrunt without a big huge effort on my pilot project and can flip over to this instead.
Mykola Lev has joined Public “Office Hours”
Andrew Thompson has joined Public “Office Hours”
Huge +1 from me and my team for taskfile.dev. We use it extensively across almost all of our projects
For those interested….. https://github.com/cloudposse/atmos/blob/master/example/stacks/workflows.yaml
This is example.
2021-07-29
| [Terraform State Security | Cloud Posse Explains](https://www.youtube.com/watch?v=xw0iPkYGKQM) |