Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

results of running audit script as of 2021-01-13 #1534

Merged
merged 39 commits into from
Feb 18, 2021

Conversation

spiffxp
Copy link
Member

@spiffxp spiffxp commented Jan 15, 2021

It's been way too long since this was last run and the results checked in. Last commit to audit/ is dated 2020-05-06. So this is ~7mo of changes

I tried to group changes into logical commits, but I won't be able to exhaustively tie each commit back to the issues/PRs that spawned it.

This is WIP because I want to either resolve or open up followup issues for commits that have QQ or TODO in their messages.

I really, really, really want us to get this running and automatically PR'ed by prow (ref: #244). Current state of affairs is broken.

specifically:
- one for each "special" pool of projects
- ingress, gpu, scale
- intended to match equivalent k8s-infra-e2e-boskos projects
changes reflected in here include:

- add k8s-infra-prow-viewers@
- add k8s-infra-prow-oncall@
- quotas updated out of band
- add ssh-keys (added by kubetest2)
specifically:
- enable secretmanager
- enable serviceusage
- reflect quota updates done out of band
- reflect automated cluster changes done out of band
specifically:
- windows-remote-docker_ca-pem
- windows-remote-docker_cert-pem
- windows-remote-docker_key-pem
specifically:
- create the service account
- bind to trusted prow via workload identity
- empower SA to deploy to prow build clusters
@k8s-ci-robot k8s-ci-robot added do-not-merge/work-in-progress Indicates that a PR should not merge because it is a work in progress. do-not-merge/invalid-commit-message Indicates that a PR should not merge because it has an invalid commit message. size/XXL Denotes a PR that changes 1000+ lines, ignoring generated files. area/audit Audit of project resources, audit followup issues, code in audit/ labels Jan 15, 2021
@k8s-ci-robot k8s-ci-robot added wg/k8s-infra approved Indicates a PR has been approved by an approver from all required OWNERS files. labels Jan 15, 2021
also drop permissions given to related kubernetes.io group
@k8s-ci-robot k8s-ci-robot added the cncf-cla: yes Indicates the PR's author has signed the CNCF CLA. label Jan 15, 2021
specifically:
- k8s-staging-addon-manager
- k8s-staging-bootkube
- k8s-staging-boskos
- k8s-staging-ci-images
- k8s-staging-cloud-provider-gcp
- k8s-staging-cluster-addons
- k8s-staging-cri-tools
- k8s-staging-etcdadm
- k8s-staging-examples
- k8s-staging-git-sync
- k8s-staging-ingress-nginx
- k8s-staging-ingressconformance
- k8s-staging-gsm-tools
- k8s-staging-kustomize
- k8s-staging-mirror
- k8s-staging-networking
- k8s-staging-provider-aws
- k8s-staging-scheduler-plugins
- k8s-staging-sig-docs
- k8s-staging-sig-storage
- k8s-staging-sp-operator
- k8s-staging-storage-migrate
specifically
- enable containeranalysis
- enable containerscanning
- enable secretmanager
- give k8s-infra-gcr-vuln-scanning SA containeranalysis viewer role
specifically:
- add k8s-conform-s390x-k8s bucket, SA for write, key in secretmanager
- add k8s-conform-inspur
specifically:
- restrict access to gsuite-groups-manager_key to serviceaccount
- allow k8s-infra-prow-build-trusted to run as SA via workload identity
@cpanato
Copy link
Member

cpanato commented Jan 15, 2021

@spiffxp I will take a look at this over the weekend

@spiffxp
Copy link
Member Author

spiffxp commented Jan 15, 2021

/uncc @justaugustus
OOO
/cc @thockin
I'm not asking that you look at this in great detail, more as an FYI. If you want to, I would recommend reviewing the commit messages and digging in on those that sound weird to you.

@k8s-ci-robot k8s-ci-robot requested review from thockin and removed request for justaugustus January 15, 2021 22:08
},
{
"members": [
"user:[email protected]"
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

All LGTM I think, it took some time to go over all the files related, but I did not see anything that calls my attention

just one comment, in several iam.json files we only see one "roles/owner" it is not better to have at least one more for backup in case it is needed? or this is expected? (asking because might be discussed before to be like this)

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Whoever runs the script that creates a project gets assigned as owner (we have a bug open about this #299). Org owners implicitly get owner on projects within the org, so we're available for fallback.

I believe the thinking goes: don't give owner access to projects, thereby requiring a gitops workflow for things like changing iam permissions. It's just unfortunate that the execution part of our gitops model is currently "humans run scripts" vs. bots.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

thanks for the clarification Aaron

@k8s-ci-robot
Copy link
Contributor

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: cpanato, spiffxp

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

@spiffxp
Copy link
Member Author

spiffxp commented Jan 20, 2021

/assign @thockin

Base automatically changed from master to main February 9, 2021 00:35
@spiffxp
Copy link
Member Author

spiffxp commented Feb 11, 2021

This is now N weeks out of date. I'd like to re-run audit and add some more commits to this, or merge this as-is and do a followup PR.

@hh is making progress on an automated CI job (kubernetes/test-infra#20742) but I'd like to get us to a smaller more digestible delta before we accept PR's from that job. If we close this PR and wait for that job to work, we're going to land something larger than this PR, but with one commit. I don't want to review that.

@spiffxp
Copy link
Member Author

spiffxp commented Feb 16, 2021

I am re-running audit.sh and will add commits to this. I'm thinking maybe it's best to leave the questionable changes listed in QQ commits out of this, to merge the non-controversial stuff here, and save discussion of those for another PR.

Copy link
Member

@thockin thockin left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I asked the same set of questions you did. I filed some issues on other weirdnesses that are not as urgent feeling.

Mostly LGTM

@@ -72,6 +72,8 @@
"group:[email protected]",
"user:[email protected]",
"user:[email protected]",
"user:[email protected]",
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We shouldn't need these (since we have group:[email protected]). But they are OK as a "safety".

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yup, "safety" is why they're explicitly defined here, is there someplace we should doc this?

],
"name": "organizations/758905017065/roles/prow.viewer",
"stage": "ALPHA",
"title": "Prow Viewer"
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I find nothing in our repo setting this up. Aaron, your fingerprint is in the access log. We should set up a script or terraform to sync the org.

#1659

@@ -0,0 +1,11 @@
{
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I wrote a quick script to diff all 30 of these.

06-30 have higher quotas than 01-05.

Not enough to abort this audit, but worth noting. Issue filed

@@ -0,0 +1,11 @@
{
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I ran a script to diff these and it was clean.

@@ -0,0 +1,11 @@
{
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I ran a script to diff these again 001.

41 + are missing the IAM binding:

    {
      "members": [
        "group:[email protected]"
      ],
      "role": "roles/viewer"
    }

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@@ -22,6 +22,32 @@
},
{
"members": [
"serviceAccount:[email protected]"
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Almost certainly what happened was that you or I or someone with "enough" privs clicked on the "compute" tab while in this project (that's my default tab) and it "helpfully" enables the service. Sigh. We should just go and disable it in the script. so it will get cleaned up on every run.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

opened #1675

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

opened #1675

@@ -20,6 +20,24 @@
],
"role": "roles/cloudbuild.serviceAgent"
},
{
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

See comment on compute. I suspect the same, but I have less evidence). We should nuke it.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

added to #1675

@@ -20,10 +20,42 @@
],
"role": "roles/cloudbuild.serviceAgent"
},
{
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

agree - should be disabled.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

given your explanation for "I clicked on console and this happens automatically" this is almost undoubtedly my fault, I navigate these services all the time on a number of projects

added to #1675

@@ -0,0 +1,8 @@
{
"displayName": "App Engine default service account",
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Maybe something to do with promoter? Can't think of a reason

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I have been in app engine before (looking at gubernator), it's possible this is the same deal as everything else in #1675

I'm surprised clicking around on console auto-activates services. Is it because our accounts are over-privileged? I'd rather see a "nope, you can't use this because the API isn't enabled for this project"

@@ -18,6 +18,12 @@
],
"role": "roles/bigquery.jobUser"
},
{
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Maybe related to promoter? @listx may know.

@thockin
Copy link
Member

thockin commented Feb 18, 2021 via email

@spiffxp spiffxp changed the title [wip] results of running audit script as of 2021-01-13 results of running audit script as of 2021-01-13 Feb 18, 2021
@k8s-ci-robot k8s-ci-robot removed the do-not-merge/work-in-progress Indicates that a PR should not merge because it is a work in progress. label Feb 18, 2021
@spiffxp
Copy link
Member Author

spiffxp commented Feb 18, 2021

Forget what I said in #1534 (comment)

I did re-run audit, but I'd rather save those for another PR. And since we've opened a bunch of followup issues for the QQ commits, I would rather merge them as-is, and we can go back and reference that these have been closed out in a later audit PR

So... looking for lgtm?

@dims
Copy link
Member

dims commented Feb 18, 2021

/lgtm

( i checked all the previous comments ... no, i did not go through 38k lines! )

@k8s-ci-robot k8s-ci-robot added the lgtm "Looks good to me", indicates that a PR is ready to be merged. label Feb 18, 2021
@k8s-ci-robot k8s-ci-robot merged commit ed90c4d into kubernetes:main Feb 18, 2021
@k8s-ci-robot k8s-ci-robot added this to the v1.21 milestone Feb 18, 2021
@spiffxp spiffxp deleted the audit-2021-01-13 branch February 18, 2021 01:30
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
approved Indicates a PR has been approved by an approver from all required OWNERS files. area/audit Audit of project resources, audit followup issues, code in audit/ cncf-cla: yes Indicates the PR's author has signed the CNCF CLA. lgtm "Looks good to me", indicates that a PR is ready to be merged. size/XXL Denotes a PR that changes 1000+ lines, ignoring generated files.
Projects
None yet
Development

Successfully merging this pull request may close these issues.

5 participants