Class EdgeIngestionEventIds

Namespace

Qavren.Edge.Ingestion

Assembly

Qavren.Edge.Ingestion.dll

Every log event Qavren.Edge.Ingestion emits, at spec section 11's numbers.

public static class EdgeIngestionEventIds

Inheritance

object

EdgeIngestionEventIds

Inherited Members

object.Equals(object)

object.Equals(object, object)

object.GetHashCode()

object.GetType()

object.MemberwiseClone()

object.ReferenceEquals(object, object)

object.ToString()

Remarks

926 (FtsMergeCompleted) is deliberately ABSENT — plan adjustment 2 deleted SP3's own FTS5 sidecar merge along with the four FtsMerge* members of IngestionOptions. The gap is left rather than closed by renumbering 927, so a log filter written against the published spec keeps matching every id that still exists.

Fields

ChunkMergedUp

A trailing under-MinTokens chunk was merged into the previous one instead of being emitted alone.

ChunkTruncated

A chunk's text was truncated to fit MaxTokens.

ChunkerFellBack

The configured chunker failed and the run fell back to a simpler one.

DocumentFailed

A document failed extraction, chunking or writing.

DocumentIndexed

A document was extracted, chunked and written successfully.

DocumentNoTextLayer

A document had no text layer (see DocumentHasNoTextLayer).

DocumentSkipped

A document was skipped — its content hash matched the last indexed run.

DocumentUnsupported

A document's extension matched no registered extractor (see ExtractorNotFound).

EmbedBatchShrunk

An embed batch was halved and retried after IngestionEmbeddingFailed.

EncodingFallback

Invalid UTF-8 fell back to Latin-1 decoding (see DocumentEncodingUndecodable).

ExtractionStreamingMode

Extraction switched to streaming mode for a large document.

ExtractionWarning

An extractor reported a non-fatal warning while extracting a document.

ExtractorSelected

Logged once per document naming which IDocumentExtractor was selected.

HeadingPathTruncated

A chunk's heading-path breadcrumb was truncated to fit HeadingPathTokenBudget.

HeadingSanitised

A heading was sanitised before being joined into a breadcrumb (see SanitizeHeading(string)).

MediImagesDropped

Embedded images were dropped from extracted content; text-only indexing continued.

OrdinalsRepaired

Chunk ordinals for a document were renumbered to close a gap left by a partial write.

PageTimedOut

A PDF page's parse time hit the page budget (see DocumentPageBudgetExceeded).

RecipeChanged

The recipe hash changed and documents were silently re-indexed under it (see IngestionRecipeChanged).

RunCompleted

A run finished successfully.

RunFailed

A run failed outright.

RunStarted

A run began.

RunSuspended

A run suspended after a failure it could not retry past (see IngestionEmbeddingFailed / IngestionRunAborted).

StaleDocumentsPruned

Stale documents were pruned at the end of a run (a source no longer yields them).

StateCommitted

A checkpoint was committed to the state store.

ThrottleAdjusted

The run's throttle interval was adjusted in response to device conditions.

ThrottlePaused

The run paused briefly under the throttle.