Export Center Observability

Audience: Operators and SREs monitoring the StellaOps Export Center service. Purpose: Describe the dashboards, metrics, logs, health probes, and alerts used to observe export runs — and how to validate them in fully offline deployments.

For end-to-end operational procedures (SLOs, capacity, failure response), see the Export Center Operations Runbook.

Dashboards (offline import)

Grafana JSON: dashboards/export-center-observability.json (import locally; no external data sources assumed).
Panels: export job duration p95/p99, bundle size histogram, registry push latency, provenance/attestation verification failures, queue depth, and error rate per profile.

Key metrics

export_job_duration_seconds_bucket{profile} — export duration by profile.
export_bundle_size_bytes_bucket{profile} — bundle size distribution.
export_registry_push_latency_seconds_bucket{profile} — registry push latency.
export_attestation_failures_total{reason} — DSSE/provenance verification failures.
export_queue_depth — pending export jobs.
export_manifest_publish_total{result} — manifest publish successes/failures.

Logs & traces

Correlate by exportId, profile, tenant; include bundleDigest, attestationStatus, registry. Traces disabled by default; enable OTLP to on-prem collector when permitted.

Health/diagnostics

/health/liveness and /health/readiness (export service) check storage, registry reachability, and attestation verification path.
/status exposes build version, commit, feature flags; verify against offline bundle manifest.
Verification probe: stella export bundle verify --manifest <path> once bundle available; validate hashes against manifest.

Alert hints

Export job duration p99 > target SLA per profile.
Attestation verification failures > 0 over 10m.
Registry push latency spikes or error rate > threshold.
Queue depth growth without completion.

Offline verification steps

Import Grafana JSON locally; point to Prometheus scrape labeled export-center.
Run stella export bundle --profile <profile> --manifest out/manifest.json and verify hashes via jq -r '.files[].sha256' against generated bundles.
Fetch /status and compare commit/version to offline bundle manifest.

Export Center Operations Runbook — SLOs, capacity, failure response.
Export Center Overview and Architecture.
Dashboard source: dashboards/export-center-observability.json.