operations view

Watch the whole pool move.

This is the operating view: where runs are going, what each vendor costs, and how policy is shaping the mix.

Chat app -> Cerver -> Compute pool

Live sandboxes

312

Monthly payment

$13.3k

Pool uptime

99.95%

Current allocation

Cerver is leaning on Cloudflare for cheaper short work, Vercel for fast starts, and E2B for deeper sessions.

Daytona on standby for canaries

Payment, uptime, start time, session length, and engagement across the live pool.

Each dot is a recent sandbox session in the pool.

X: faster start -> longer session Y: cheaper -> higher payment

cheap

higher payment

fast start

longer session

Vercel

Cloudflare

E2B

Change routing once. Let the pool move underneath it.

Default short sessions to Cloudflare Keep the cheapest, high-uptime cluster as the default route for short work.

Escalate fast starts to Vercel Use Vercel when launch speed matters more than blended pool cost.

Keep deep sessions on E2B Longer and more engaged sessions stay on E2B once they cross the pool threshold.

Chat with policy example

Move more long sessions from Cloudflare to E2B after 15 minutes.

Route through Cerver once, then steer the whole pool from one policy layer instead of rewriting each vendor integration.