Anavsan Explore
All terms Start Free Trial

Terminology › APEX loop

What is a Private Knowledge Graph?

A Private Knowledge Graph is why Anavsan can assign a cost issue to your engineer instead of a generic persona. It learns your estate. It does not treat your workloads as the industry average.

A Private Knowledge Graph (PKG) is an environment-specific map of Snowflake users, workloads, query patterns, ownership signals, Git context, and fix history. It is the memory layer under APEX: without it, Trace can still show a warehouse, but Assign cannot name a party you will trust.

Key takeaways

Private means your environment. The graph is assembled from your metadata, not a public corpus of Snowflake blogs.
It is substrate, not a chatbot. The PKG feeds assignment, simulation context, and closure history. Ask APEX reads it; it is not the product by itself.
Metadata only. The graph is built from ACCOUNT_USAGE and related signals. Business tables do not leave Snowflake.
Time is the feature. Slow drifts — a warehouse that creeps up after each migration — show up only with months of context, not a 48-hour anomaly window.

Why generic models cannot assign your issues

Industry-average “right-size this warehouse” advice ignores your coding patterns, your BI cluster’s shared schedule, and which engineer fixed this class of regression two quarters ago. Wrong assignment is worse than no assignment: you accuse the wrong team of wasting money. The PKG exists so routing uses your identity map, your query signatures, and your fix history.

That is also why workload governance is a third-party layer. Snowflake’s native views are necessary telemetry. They are not a memory of who owns what across billing cycles.

What the graph maps

  • Snowflake users, roles, and service accounts to teams and, when declared, cost centres
  • Workloads: recurring queries, pipelines, dashboards, Cortex functions
  • Cost signatures: idle, spill, scan, refresh storms, storage retention
  • Ownership evidence: tags, roles, usage hints — with confidence that never promotes a guess to a personal accusation
  • Fix history: what changed, when, and whether credits moved

Anavsan’s PKG ingests 200+ Snowflake signals across queries, warehouses, storage, Cortex, serverless compute, and related logs. Continuous sweeps keep the graph current; assignment quality improves as history accumulates (typically weeks, not minutes).

What it is not

Not thisPKG
Data graphA catalog of customer recordsA map of metadata about how you use Snowflake
LLM memoryChat history pasted into a promptStructured links among objects, people, and outcomes
CMDB cloneA static inventory you maintain by handAssembled from Snowflake signals and updated on a cadence

How it is used

Trace uses the graph to attach a workload to credits. Assign uses it to pick a channel: high-confidence declared owners can reach a person; inferred ownership stays at team level. Prove uses prior similar fixes as context. Enforce uses the trail so a returning pattern is not a brand-new alert. Product behavior is documented on APEX.

APEX

The engine that builds and uses the PKG.

Read →

Trace → Assign → Prove → Enforce

The loop the graph makes specific.

Read →

Cost Accountability

The outcome: a named owner, not a guessed one.

Read →

Enforcement Desk

Where PKG-backed findings show up as rows.

Read →

Frequently asked questions

An environment-specific map of Snowflake users, workloads, ownership signals, and fix history. It is built from metadata so assignment and proof are specific to your estate.
No. It is assembled from Snowflake metadata such as ACCOUNT_USAGE. Business records do not leave the account.
Your workloads are not the industry average. Generic models cannot remember which engineer fixed which class of issue in your Git history or which warehouse is the shared BI cluster.
It begins assembling as soon as metadata is connected. Longitudinal assignment quality typically improves over weeks as patterns and ownership evidence accumulate.