Settings

Pipeline configuration

GodEye Callback

Enabled

After each /ingest analysis, results will be POSTed as JSON to this URL.

A1 Isolation Forest (log-ml)

Enabled

AI (LLM) — used by MiroFish · Synthesizer · Perplexica

Enable LLM enrichment

Set one Default AI below; every AI stage uses it unless you turn on its Override.

Default AI

MiroFish (A3) — enabled

เปิด = ให้ 5 expert frames เรียก LLM แล้วเอาความเห็นไปรวมวิเคราะห์ (root cause). ปิด = ใช้ deterministic frame scoring อย่างเดียว (เร็วกว่า, ประหยัด LLM calls).

MiroFish (A3) — override

Using Default AI (ollama · gemma4:e4b)

Synthesizer (AA) — override

Using Default AI (ollama · gemma4:e4b)

Perplexica (A2) — override

Embeddings always stay on the local Ollama/Transformers model; this only switches the chat model.

Using Default AI (ollama · gemma4:e4b)

A2 Perplexica (External Knowledge)

Enabled

Chat model/provider for A2 is controlled in the AI (LLM) section above (Perplexica stage).

Topology (network + service) — v2

Upload a topology snapshot JSON from GodEye: { nodes: [...], edges: [...] }. Network and service layers can be uploaded as separate files — edges are replaced per layer, nodes are merged.

No topology stored for this tenant yet.