Tag: fleet engineering

Investigation and resolution of recurring Inoreader daily limit incidents.

A new Cultural Radar feature led to a surge of ‘recency_unverifiable’ incidents due to an unhandled timestamp format.

Addressed false-positive ‘down’ alerts for webhooks and completed the deprecation of ‘DaaS’ naming conventions.

A series of ‘all_engines_failed’ errors across orchestrator, vision, and fact_checker services were traced to an incorrect service name configuration.

An increase in ‘fact_check_failed:last_resort’ incidents was traced to a content mix change and an underlying model rebalance.