Cleetus Core developers

Core developers

Execution observability

The live execution graph shows what a Core can observe about its current runs. An operation receipt records a past decision or effect. Use the graph for “what is running here?” and the receipt for “what happened to this request?”

From enforcement to an observation
Kernel or VMEnforces the admitted boundary
Runtime ownerBinds measurements to this run
Graph readerDisplays the observation until it expires

Arrows show the direction of evidence. A graph reader cannot enforce a limit or grant permission to execute.

Inspect a live run

cleetus-host inspect execution-graph

Local graph inspection requires a current graph grant. Read the source Host, runtime instance, observation time and expiry together. The graph can include sessions, run generations, containment boundaries and measurements supplied by their respective owners. Measured cgroup CPU usage and memory limits describe a particular cgroup. A provider name in a catalog supplies no such measurements.

To display a remote graph, review a route at each end, then request the exact exporting Host and instance. Drop expired snapshots from the live display when a route closes. Preserve the distinction between a fresh empty graph and an unavailable source.

Recover an operation

cleetus-host inspect operation "$OPERATION_ID"

For a managed launch or control operation, this reads the stored launch state and hold, the latest control action and its result. After a restart, distinguish the original instance from the current reader. A leader exit alone gives no evidence that every descendant stopped.

Share organization metadata

Optional organization reporting has separate permission, reporting identity, workspace enrollment and reader scope. The preview carries bounded metadata, such as turn status, permission outcomes and partial usage. The live execution graph has its own disclosure rules. Organization controls · Audit and original receipts