Clarify hosted-agent vCPU concurrency limits - #3161
Conversation
Amp-Thread-ID: https://ampcode.com/threads/T-01a06a24-5e51-700f-af80-b14614169400 Co-authored-by: Daniel Ring <daniel.r@buildkite.com>
There was a problem hiding this comment.
Want to dig deeper?
Paste this into your agent to explore the findings from this review's Buildkite build:
Download the buildsworth logs from build 25405, then answer my questions about the findings.
Install the reading-buildsworth-logs skill to run this.
About buildsworth
Model: gpt-5.6-sol with xhigh thinking.
How to request a review: Comment @buildsworth-bk review on the PR, or request buildsworth-bk as a reviewer.
Risk labels (how buildsworth classifies risk) — buildsworth classifies risk itself from the diff. Unless repository policy already allows L2 approval, grant it by mentioning @buildsworth-bk (see approval ceiling and L2 approval grant):
- L1 — Low risk (dep bumps, docs/copy, lockfiles, small presentational fixes). buildsworth may approve by default.
- L2 — Standard risk (new UI, additive API fields, refactors). Approved only when repository policy or a verified grant allows it; otherwise comment-only.
- L3 — High risk (auth, migrations, payments, secrets, perf-critical pipeline paths). Human review always required.
| @@ -2,8 +2,8 @@ | |||
|
|
|||
| This page outlines usage limits designed to protect your builds from unintentional resource issues and ensure reliable service for all customers. Limits vary by subscription tier: | |||
There was a problem hiding this comment.
Non-blocking: The All Access Trial is an evaluation period rather than a subscription tier, so this lead-in misclassifies it.
| This page outlines usage limits designed to protect your builds from unintentional resource issues and ensure reliable service for all customers. Limits vary by subscription tier: | |
| This page outlines usage limits designed to protect your builds from unintentional resource issues and ensure reliable service for all customers. Limits vary across plans and the All Access Trial: |
Amp-Thread-ID: https://ampcode.com/threads/T-01a06a24-5e51-700f-af80-b14614169400 Co-authored-by: Daniel Ring <daniel.r@buildkite.com>
Amp-Thread-ID: https://ampcode.com/threads/T-01a06a24-5e51-700f-af80-b14614169400 Co-authored-by: Daniel Ring <daniel.r@buildkite.com>
|
Closing as a duplicate and superseded by https://github.com/buildkite/docs-private/pull/2181 |
An updated duplicate of buildkite#3161 Summary - Update hosted-agent concurrency limits to use combined vCPU capacity. - Document current Free and Pro concurrency limits with machine-size examples. - Restore included Linux vCPU minutes and clarify M4 Mac usage pricing. - Preserve hosted-agent cache storage quotas. - Replace outdated Personal and Trial plan terminology. - Direct trial and custom-limit organizations to Service Quotas. Validation - Vale - markdownlint - Alex - 80 page-rendering specs Changes from the original PR - Restored explicit monthly usage information: - Free: up to 2,000 Linux vCPU minutes. - Pro: 4,000 included Linux vCPU minutes. - Pro M4 Mac: pay as you go. - Enterprise: volume discounts. - All Access Trial: refer to Service Quotas. - Changed the introduction from “subscription tiers” to “plans and the All Access Trial,” because the trial is not a plan. - Reordered the list so Free, Pro, and Enterprise are grouped as plans, followed by the trial. - Updated the pricing paragraph because allowances are now documented directly rather than only linked to Pricing. Everything else from PR buildkite#3161 was retained.
Why
The hosted-agent limits table described concurrency as machine counts and used outdated plan names and values. Because each instance shape consumes a different number of vCPUs, this did not answer how many jobs an organization can run concurrently.
What
Sources were the current public pricing and hosted-agent documentation, cross-checked against current
bk/bkentitlement, shape, and service-quota configuration.billing/foundationwas not accessible from this workspace, so no internal-only billing details are included.Validated with Vale, markdownlint, Alex, link HTTP checks, and the 80-example page-rendering spec.