[ PROOFECONOMY ]™GlossaryOriginSpecsBlack Hole
A black hole is any period during which part of the system under test disappears for reasons unrelated to any canary or attack, such as a network, node, Effect Point, observer or Time Anchor outage. Canaries released into one are voided: never scored and never backfilled.
What it means
A dead canary inside a black hole is not a stop. If the Effect Point was down, nothing could have landed there, and the control gets no credit for it.
The rules
- Detected independently of the control: liveness checks on every Effect Point, the Ingress Observer and the Proof Witness over paths that do not cross the control; good canary loss across both labels at once; gaps in observer or witness records; missed Time Anchor bindings. The Liveness Prober (IP-7) does this work.
- Recorded as a window: start and end bound to Time Anchor pulses, the affected components, the detection method and the evidence of the cause.
- Voided: every canary released into, or landing inside, the window is adjudicated
void_black_holeand excluded from all rates. A run whose void share exceeds its declared threshold is invalid. - No backfill: voided canaries are never replaced by re-labeling, re-timing or reusing their commitments.
- Witness black holes: while the Proof Witness is dark there is no proof, and those windows are published.
- Black Hole Rate: the share of time and of canaries lost to black holes is published with every result series.
What it is not
An outage caused by the control is not a black hole. If the control fails closed, crashes or drops traffic, the effect on good canaries is a False block and is scored as one. A control that breaks the path is failing, not absent.