WHITEBOARD / BOARD: TEAM POOLS
Board: team pools and cost
The goal is to show a team sharing capacity, where adding people costs nothing and adding CPU costs by the hour.
Whiteboard outcome
When you’re done, the board shows a team (owner, admin, members), two pools in different regions, VMs inside them, the placement rule, and the monthly bill written next to each pool.
Before you draw: three questions
- How many people, and do they trust each other’s VMs?
- Is the workload steady or bursty (a CI week)?
- Which regions do your users and data sit in?
Draw it in 8 steps
- At the top left, draw a team box with three stick figures: Owner, Admin, Member. Say: “Work is $75 a month for the whole team, not per person” (https://exe.dev/docs/billing/work).
- On the right, draw a dashed rectangle “pool: build, 8 vCPU / 16 GiB, LAX” with three small VMs. Say: “A pool is reserved CPU and memory in one region. It’s billed while it exists, empty or full” (https://exe.dev/docs/pools).
- Below it, draw a second dashed rectangle “pool: prod, 4 vCPU / 8 GiB, FRA”. Say: “On Work you can have unlimited pools, in any region” (https://exe.dev/pricing).
- Draw an arrow from the members to the build pool labeled
team settings vm-placement pool build. Say: “newlands in the default pool, and--poolalways overrides it” (https://exe.dev/docs/pools). - Write the cost next to each pool, with the sum below. Each pool’s 2 vCPU is included, and the rest is metered. Say it as derived math: the build pool is 3 extra units × $0.105 × 730.5 h ≈ $230.11, and prod is 1 extra unit ≈ $76.70. See Economics for the rules [verify: how the included 2 vCPU is applied across several pools].
- Draw a small calendar with one week shaded, labeled
pool resize build --cpus=16, then shrink it back. Say: “If you grow a pool for a busy week, the extra costs you that week only” (https://exe.dev/docs/billing/work). - Draw a padlock on the team box labeled “SSO + vm-sharing admins-only”. Say: “Use OIDC or Google sign-in, and only admins can share outside the team” (https://exe.dev/docs/teams/sso, https://exe.dev/docs/teams/sharing-controls).
- At the bottom, write the rule: “Pay for capacity, not seats. Resize, don’t over-buy.”
flowchart LR
subgraph T[Team on Work, $75/mo base]
O[Owner]; A[Admin]; M[Members]
end
M -->|new| B
subgraph B[pool build: 8 vCPU / 16 GiB, LAX]
b1[ci-1]; b2[ci-2]; b3[dev]
end
subgraph P[pool prod: 4 vCPU / 8 GiB, FRA]
p1[api]
end
A -->|pool new / resize| B
A -->|pool new| PThe included-capacity detail, said carefully
- Included capacity. The Work plan includes 2 vCPUs and 4 GiB “across all of the team’s pools and standalone VMs”, applied to whatever is running at the same time (https://exe.dev/docs/billing/work, https://exe.dev/docs/standalone-vms).
- The pools doc example. “A team with one 4 vCPU pool pays the monthly price plus $0.105 per hour for the 2 vCPU above the included capacity” (https://exe.dev/docs/pools).
- Two pools. The included capacity is applied once, not once per pool. On the board, subtract it once from the total.
What they’ll ask
| Question | Answer | Source |
|---|---|---|
| “Do more VMs cost more?” | No. Capacity costs money, not VM count. Each pool holds up to 1,000 VMs. | https://exe.dev/docs/pools |
| “Bigger than 32 vCPU?” | The pool waits for approval, and a pending pool isn’t billed. | https://exe.dev/docs/pools |
| “Can admins get into my VM?” | Yes. Team admins can SSH into members’ VMs. | https://exe.dev/docs/teams/vms |
| “Untrusted members?” | team settings core-sharing off gives each VM its own physical core. | https://exe.dev/docs/teams/vms |