@here we are in the process of forking Atlantis to keep a more active project and with more maintainers to make the project better and add some needed features, what you will like to see merged from existing PRs and or new features?
- Hiding previous comments
- Allowing default plan and custom plans - they seem mutually exclusive from my testing
- Allowlist select users or groups to do applies
- More logging
can you expand on Custom plans? and more logging(server side or repo side)?
API to trigger ad-hoc plan/apply
Hiding previous comments < this is already a thing.
@Erik Osterman (Cloud Posse) mentioned that CloudPosse moved away from Atlantis due to some highly-desired features not making the cut. He might have some ideas
Please do announce us of when you’re forking it and what your plans are!
I will make sure to keep you guys updated
Erik already gave me a list of things of what he will like to see
Ability to run a custom script that generate
atlantis.yaml dynamically instead of reading it from within the triggering repo. The script would have access to the cloned repo so it can generate the configuration based on the repo folder structure and other values in the repo.
you mean like generating a default config base on the repo structure?
Here’s an example based on the old geodesic folder structure:
If we could make arbitrary scripts that ran whenever there was a webhook event, thats even better. Another potential application is a script that clones external repositories and merges it into the triggered repo so that “common” projects could be pulled in.
That is interesting, but it will be a bit tricky
Atlantis clones the repo after and event is trigger so a setting of this type will have to be an the server config level
Mmmm is actually not that complicated, and it could support pulling the
common folder from VCS or S3 ot local
Generate Atlantis config for Terragrunt projects. Contribute to transcend-io/terragrunt-atlantis-config development by creating an account on GitHub.
thanks for doing this. Will try to pitch in when i can
you can already hide previous comments?
Problem I have ATLANTIS_HIDE_PREV_PLAN_COMMENTS=true and running 0.15.0 as a github app in ECS Fargate. I run plan manually in my repo using the following. atlantis plan -d iam-policies/engineer Wh…
feel free to help with that by commenting the thread. a lot of people are stuck on it. im also using atlantis as a gh app.
i am using gitlab, i should give this a try
that is interesting since It works for me but I’m not on 0.15.0
but I’m using webhooks, which I do not believe is related to the issue
this is in pair with master if a few weeks back
@PePe where do you set the flag on the server config?
I actually did not set it
i am using the upstream terraform module with fargate
I start atlantis like this :
dlv debug --headless --listen=:2345 --log --api-version=2 \ ~/workspace/Go/src/github.com/runatlantis/atlantis/main.go -- \ server --atlantis-url="<https://73c90a215719.ngrok.io>" --gh-user="jamengual" \ --gh-token="wretwertwertwertwerwe" --gh-webhook-secret="ertwerwertwert" \ --repo-allowlist="github.com/jamengual/atlantis_testing" --log-level="debug" \ --config=atlantis-repo-config-staging.yaml
weird and your old comments get deleted on a new plan?
they do sometimes
I can’t really tell if they do since this is my testing branch but in general they seem to get hidden not deleted
I see this functionality working as expected…
it looks like this is only available in github @joshmyers
I have been added as a Maintainer on the Atlantis project and we are now working trough PRs and merging some and there is some cool features coming soon
Do you have a TL;DR on new features?
there is a bunch of merged pr you can look at and we are reviewing a few more
we will be cutting a release late this week or next week
we are new as contributors so we are automating a few things so that we can iterate faster
there is one new PR that we are looking at that will stream the terraform run on the atlantis UI so you can see what is happening
the Github Teams PR needs test which currently I’m working on it
there has been a lot of work on Workflow customization lately
repo side and server side
streaming the tf apply output somewhere would be epic
this has just started happening for us, does anyone have any ideas?
running "/usr/local/bin/terraform init -input=false -no-color -upgrade" in "/home/atlantis/.atlantis/repos/ume-platform-engineering/tf-organisation/9/default": exit status 1 Error: Unsupported Terraform Core version on config.tf line 10, in terraform: 10: required_version = "= 0.13.4" This configuration does not support Terraform version 0.13.3. To proceed, either choose another supported Terraform version or update this version constraint. Version constraints are normally set for good reason, so updating the constraint may lead to other errors or unexpected behavior.
Yeah, you have a
required_version = "= 0.13.4" constraint somewhere and you are trying to use 0.13.3 with it….
the constraint is there
and the image we are using is v0.15.1
If you’re using the Atlantis Docker image and aren’t setting the
then the default version of Terraform will now be 0.13.4. Simply set the above
flag to your desired default version to avoid any issues.
i might just explicitly set that verison
we’re using tfenv inside Atlantis to manage versions for us so that Atlantis uses the same path as we do if working with the codebase locally ¯_(ツ)_/¯
i managed to get atlantis to plan but now on the apply i am getting …
stat /home/atlantis/.atlantis/repos/ume-platform-engineering/tf-organisation/9: no such file or directory
the atlantis.yml in the repo is
version: 3 projects: - name: latest dir: . autoplan: when_modified: ["*.tf"] enabled: true
this has been working fine for ages i think its because the upstream tf module uses the latest image by default
Yeah, don’t do that.
the fact they default to latest is crazy imo
what i don’t know now though is why the apply errors out
any ideas as i am stumped
Where do you see that stat log line?
os.Stat is called in a few places in Atlantis codebase
I’m assuming it is complaining about the PLANFILE not being present for the apply after a plan?
weird after doing a new plan and apply it worked
I was going to say that is you close and open the pr it will clean the cloned dir in the Atlantis server and that way you could fix that error
But a new plan does a similar thing
Anyone here using the Slack Atlantis webhook integration?
I’m getting the same as https://github.com/runatlantis/atlantis/issues/1210
Hi, We integrated slack to atlantis with following config in config.yaml file slack-token: xxxxxxxxx webhooks: event: apply workspace-regex: .* kind: slack channel: test and when starting atlantis …
Bah, looks like Slack have deprecated methods in their API and slack-go hasn’t been updated yet (and Atlantis will be using an old version)
this is my next thing to setup
It doesn’t work natively via Atlantis due to Slack API deprecations
i am using the upstream tf module as well and don’t see any integration points either
Nope, you need to add some webhook config to the Atlantis image which is undocumented
is there another go slack library?
Probably, but it isn’t the one Atlantis is using.
Everyone is running TF through some wrapper anyway, right
Atlantis is falling behind sadly