Discloud automatic restart delay for crashed processes across plan tiers

Discloud documents automatic restarts for crashed service processes, but the precise restart delay and backoff behavior are not fixed in the available documentation. The goal is to determine whether the delay is constant, progressive, or configurable, and whether it varies between deployment paths or service tiers.

Free and paid plans differ in concurrent instances, memory, and CPU, creating a compatibility boundary that may affect restart handling. Behavior could also diverge between the web dashboard and CLI workflows. Current CLI names, manifest schema, and exact tier limits are unverified, so these details require confirmation against the live platform.

What is the default restart delay after a Discloud service process crashes?

Does the restart backoff increase after repeated crashes, and can it be changed in service settings?

Does restart timing differ between free and paid tiers or between dashboard and CLI deployments?