Excititor Chunk Telemetry

Implementer: Excititor Observability Guild
Reviewers: Evidence Locker Guild (for parity with attestation metrics)

Reference: Sprint 110.

Audience: operators and observability engineers monitoring the Excititor Chunk API.

This page documents the metrics and logs emitted by chunk ingestion. For the broader set of Excititor evidence metrics, see the Observability guide.

Metrics

All instruments are published under the StellaOps.Excititor.Chunks meter.

Metric	Type	Unit	Description
`vex_chunks_ingested_total`	Counter	chunks	Increments per chunk submitted. Tags: `tenant`, `source`, `status` (`accepted`/`rejected`), `reason` (empty for accepted).
`vex_chunks_item_count`	Histogram	items	Item count per chunk.
`vex_chunks_payload_bytes`	Histogram	bytes	Measured from the NDJSON payload length.
`vex_chunks_latency_ms`	Histogram	ms	End-to-end ingestion latency per request.

Event	Fields
`vex.chunk.ingest.accepted`	`chunk_id`, `tenant`, `source`, `item_count`, `chunk_digest`
`vex.chunk.ingest.rejected`	`chunk_id`, `tenant`, `source`, `reason`, summarized validation errors

Register ChunkTelemetry as a singleton bound to the shared Meter instance.
In the /vex/evidence/chunks handler, compute chunk_digest deterministically from canonical JSON and emit the counters/histograms via ChunkTelemetry.
Log using the structured templates above; never include request bodies in logs.
Expose metrics through the default ASP.NET metrics export (Prometheus/OpenTelemetry) already configured in the WebService.

Do not include host-specific paths or timestamps in metric dimensions.
Use the standard OTEL default histogram buckets; do not generate buckets at runtime.
Keep the meter name stable. Adding new instruments requires a version note in the sprint Decisions & Risks.