nilalytics v0.3.0

Highlights

• Subject partition + person bucketing for user_events. The curated table is now laid out subject › date › bucket(person_id) and sorted by person_id, so category, time-window, and single-person reads all prune well. person_id is hashed into N buckets (bucket(N, person_id)) — DuckLake's recommended way to partition a high-cardinality key without the small-files problem. Tune with NILA_USER_EVENTS_BUCKETS (default 16).
• Subjects. Every event is classified into a low-cardinality subject — errors / activities / ai_usage / traceability / other. An explicit nila.subject attribute wins; otherwise it's derived (severity/exception → errors, gen_ai/llm → ai_usage, identify/audit → traceability). Unknown values clamp to other.
• Identity lookup by any key. Users are defined by any stable, unique key (email / account id / phone), hashed into a person_id. New nilalytics identify <key> prints that id, and nilalytics query user --key <value> [days] hashes it for you.
• New reads. query subject <name> [days], a subject breakdown in query user_events, and the emitter now sends sample ai_usage events.

Upgrade note

The user_events table gains a subject column and a new partition layout. On first start after upgrade, the server rebuilds it from otlp_logs (the source of truth) automatically.

Full changelog: v0.2.0...v0.3.0

nilalytics v0.2.0

Highlights

• Curated user_events table (recommendations / ML). Identity and the common event fields are lifted out of JSON into typed columns, partitioned by day and sorted by person_id, so "one user's history" is a fast, file-pruned read. Refreshed incrementally via the DuckLake change feed (snapshot-id watermark → exactly-once, order-safe).
• Per-user activity view. nilalytics query user <id> [days] returns one person's full activity and logs, optionally windowed (e.g. last 16 days), with a per-event-type breakdown.
• Opt-in data retention. NILA_RETENTION_DAYS drops event rows older than the cutoff (across otlp_logs, otlp_traces, otlp_metrics_*, user_events) so storage stays bounded. Runs inside the single-writer server. Distinct from NILA_RETENTION_MS (snapshot/time-travel retention).
• Retention dry-run. nilalytics maintenance --retention-dry-run [--days N] previews exactly what retention would delete, per table — read-only, deletes nothing.
• Backend (server-side) activity guide. Documents backends as a first-class OTLP source and how to tie backend spans to the same user/action.

Docs

New User events and Backend activity guides; expanded configuration, querying, maintenance, and architecture pages.

Full changelog: v0.1.0...v0.2.0

nilalytics v0.1.0

First release of nilalytics — serverless realtime web & mobile analytics on
DuckLake + object storage (S3/GCS/R2/Azure), served over DuckDB's Quack protocol.

Highlights:

• OTLP ingest (duckdb-otlp) into DuckLake with data inlining
• Public ingest gateway: CORS, short-lived tokens, optional TLS
• Multi-cloud object storage, read-only Quack query authz
• Cross-device pseudonymous identity, funnels/retention/traces/metrics
• nilalytics CLI + full docs at https://angelerator.github.io/nilalytics/

Install: pip install nilalytics