Observe the slope
Start containers at quartile levels, measuring time-to-running while published bindings and mounts accumulate.
Preflight survey / Docker + Podman
Measure startup p95, published-port pressure, mounts, and headroom on the host you actually plan to use—then validate the prediction with a second controlled run.
Field evidence / 01
The probe turns a sandbox shape into host-specific evidence instead of a generic container limit. Every result names the runtime, context, image, load shape, and caveats.
Start containers at quartile levels, measuring time-to-running while published bindings and mounts accumulate.
Fit a conservative p95 trend and compare it with your startup budget: comfortable, watch, or exceeded.
Repeat the same shape later. scp compare reports prediction error and whether it meets the 25% target.
Runbook / 02
$ cargo install --path cli $ scp probe \ --target staging-west \ --confirm staging-west \ --containers 12 \ --ports-per-container 2 \ --mounts 1 \ --samples 3 \ --output capacity.json Sandbox capacity envelope planned shape 12 × 2 ports × 1 mount level starts p50 p95 published 3 3 182ms 194ms 6 6 3 219ms 238ms 12 9 3 271ms 294ms 18 12 3 326ms 351ms 24 COMFORTABLE — predicted p95 351ms Next: repeat, then run scp compare.
Route sketch / 03
This local planner is a transparent planning estimate, not a benchmark. Use the generated command to replace it with evidence from your runtime.
Estimated p95 / model only
330 ms
comfortablePlanning estimate: 330 milliseconds p95, comfortable, with 1170 milliseconds headroom.
scp probe --target staging-west --confirm staging-west --containers 12 --ports-per-container 2 --mounts 1 --samples 3 --startup-budget-ms 1500 --output capacity.jsonOperator pack / 04
Planner Pro saves and compares up to five local scenarios. The complete CLI, safety behavior, JSON reports, and CSV export remain free.
$39 one-time purchase
Sociobot/Dodo is the merchant of record. Refunds are handled there and revoke the license automatically. See privacy and terms.