What problem does it solve?
Answering "what was this interface doing last Tuesday" from Zabbix is error-prone: the API silently returns empty arrays when you query with the wrong value type, mix value types in one call, or request a window beyond raw retention. This Skill enforces the correct procedure so historical answers from the NMS are accurate instead of confidently wrong.
Core Features & Use Cases
- Correct history retrieval: Always runs item.get first to learn each item's value_type, retention windows, and units before calling history.get with the matching type.
- History vs trends routing: Routes queries between raw polled values and hourly min/avg/max aggregates based on the requested time window, and discloses which source was used.
- Empty-result diagnosis: Distinguishes the five causes of an empty response (wrong value type, aged-out data, disabled retention, never-collected items, genuinely idle metrics) so "no data" is never misreported.
- Use Case: An engineer asks whether a WAN link was saturated overnight. The Skill queries the interface items with their real value types, merges raw history with hourly trends across the retention boundary, and reports peak utilization with units, timezone, and source clearly stated.
Quick Start
Ask the agent to show what the traffic on a specific switch interface looked like over the past week using Zabbix history.
