Storage Analysis for Billing

[hh]

We need to have more detailed reporting on storage, so we can detail our spend per project.

I'm creating this tracking sub-issue for #156
/cc @justinsb

[spiffxp]

Per bucket? Per path?

Costs we think of are:

• storage at rest
• bandwidth delivered (likely the larger of the two)

Figure out bucket structure that makes sense for us from a billing perspective

GCR doesn't give us this breakdown, but maybe GCS does

Use test data? Simulated load?

[spiffxp]

/assign @justinsb @thockin

[spiffxp]

Status quo

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

[thockin]

/lifecycle frozen
/remove-lifecycle stale
/remove-lifecycle rotten

[justinsb]

Just to document some status here:

We are starting to upload some artifacts, so we are starting to see some GCS costs. However these costs are likely going to be much smaller than the transfer costs once we start serving binaries. I added a page to the billing report that just breaks down GCS by project, which actually lines up pretty closely for now and shows that the numbers aren't that big. We might just not need this report.

Once we get some data transfer going, then I have a tool to build a report for S3 transfers, so I can probably do something similar for GCS.

[spiffxp]

Would be interested in seeing the data. Not super important but may give visibility into what images are high traffic etc.

[spiffxp]

We tried turning this on a while back and had to turn it off for reasons I can't recall. May be able to turn it on again soon?

[spiffxp]

/assign @hh
since you had expressed interest in working on this

[hh]

@cncf legal should have an updated @ii Contract soon to included the appropriate PII clauses.

I'll pick this up once we have that signed, and can add our team to the appropriate google group + perms.

@spiffxp does the appropriate group with perms exist yet or should we work on that as a pair/debug session to ensure principal of least privilege.

[spiffxp]

We likely have work to do.

#1726 might help in putting together the right role

But identifying the right resources to apply it to may take some time. I'm pretty slammed but I'll see if I can find some time next week too follow up with you off PR

[riaankleinhans]

/assign @Riaankl

[spiffxp]

I think this is no longer blocked, given that @Riaankl has been doing analysis of images transferred from k8s.gcr.io

[spiffxp]

/milestone v1.23

[spiffxp]

Mildly related: #2710 attempts to figure out where images are going

Not necessarily how much each is costing us

[ameukam]

/milestone clear