Module metrics 

Prometheus surface for SLA monitoring (#202).

• faucet_pipeline_sla_violations_total{pipeline,row,kind} — counter; kind ∈ staleness | min_rows | volume.
• faucet_pipeline_sla_baseline_runs{pipeline,row} — gauge; successful runs currently in the rolling volume baseline (cold-start visibility).

Emission follows the CLI-side convention (faucet_schedule_*, faucet_serve_*): plain metrics macros against whatever recorder install_observability installed. Labels stay low-cardinality (pipeline

• row only, like every pipeline metric — never record keys or run ids).

Functions

record_violation

Count one detected violation.

set_baseline_runs

Publish the baseline depth after a successful run folds in.