proof-server:proof-server-operations

A 503 response indicates proof server queue saturation or readiness failure. Inspect the /ready endpoint to check jobsProcessing and jobsPending metrics, then review log patterns for job failures to determine if you need to increase workers or scale horizontally.

Proof server readiness is checked by querying the /ready and /health endpoints. You can configure Kubernetes readiness probes to poll these endpoints, interpreting job queue metrics and worker utilization to determine liveness and capacity saturation.

Scaling proof servers requires tuning --num-workers and --job-capacity based on k-value and worker count metrics. Estimate throughput capacity from queue utilization data, then apply horizontal scaling or adjust worker parameters to resolve saturation.

Yes, Docker health checks support Midnight proof server deployments. You can configure Docker probes to monitor the /health and /ready endpoints, tracking jobsProcessing and jobsPending metrics to detect queue saturation and trigger automated restarts.

Troubleshoot proof server timeouts and memory issues by analyzing log patterns for error signals like 503, 429, and 400 responses. Identify job failure and cancellation entries to isolate capacity bottlenecks and determine remediation steps.

Increase --num-workers when proof server job capacity allows more concurrent processing, but scale horizontally when worker utilization is saturated and throughput gains plateau. Compare jobsPending against k-value limits to decide the optimal scaling approach.