Full Observability and Auditing

Features

1. 

   05.1

   One Run ID for Every Execution

   Endpoint calls, schedules, triggers, record processing and queued previews each produce a run. Every run has a complete, never-sampled step log.

   • Find every execution in one list.
   • See every step of every run.
   • Read complete logs, never samples.
   • Inspect scheduled and triggered runs.
   • Debug previews like production runs.
2. 

   05.2

   What Each Run Wrote

   Every run records its writes as applied, rejected or discarded. Writes apply as one set, so a failed run leaves no half-written records.

   • See what every run wrote.
   • Spot rejected and discarded writes.
   • Audit changes run by run.
   • Never clean up half-written records.
3. 

   05.3

   The Exact Version That Ran

   Each run keeps a snapshot of the flow it executed. Project history lists every configuration change with every row it touched.

   • See the exact flow that ran.
   • Compare a run with today’s flow.
   • See every configuration change.
   • See which rows a change touched.
   • Answer "what changed, and when".
4. 

   05.4

   Warnings That Tell Cases Apart

   “Could not look” reads differently from “looked and found nothing”. A fault on the platform’s own provider account is labelled and not charged.

   • Tell failures from empty results.
   • Know when a source couldn’t answer.
   • See platform faults, labelled clearly.
   • Never pay for platform faults.
5. 

   05.5

   Refused and Model Calls on Record

   Every request a product endpoint refused is logged with its reason. Every model call is listed with its outcome.

   • See every refused request’s reason.
   • Debug failing client integrations.
   • List every model call made.
   • Check each model call’s outcome.
   • Review model costs call by call.