Guides
How ingest and correlation work
XFlow is the operator dashboard for your app ecosystem. It shows what is connected, what is broken, what needs attention, and what actions keep your apps running. Technically, that requires a single ingest contract, explicit validation outcomes, and correlation identifiers that stitch retries and parallel work into one narrative.
Entry layer (operator language)
Operators need to answer: what changed, for whom, and with what evidence. Ingest exists so those answers are not reconstructed from screenshots and memory.
Pipeline (engine language)
Ingest
Events arrive through a typed HTTP surface. The system validates payloads and returns explicit outcomes—created, skipped, or rejected—rather than silent success.
Idempotency and deduplication
Retries are inevitable. Idempotency keys and dedupe rules exist so the same logical event does not become multiple incident signals.
Normalization
Records are normalized to a canonical shape with environment and app tags so different emitters can be triaged consistently.
Correlation
Request IDs and related keys tie retries, parallel deliveries, and cross-app hops into one explainable story for triage.
Evidence and contracts
Deeper technical sources live in the repository (event model and API contracts). Start from the docs index for paths.
Next steps
Move from reading to evaluation: see authenticated surfaces, redacted walkthroughs, or the operator capability map.