Backup & restore

novamem's source of truth is Postgres. The cold-tier (Qdrant) and graph-tier (FalkorDB) are reconstructible from the warm-tier embeddings and edge auto-link logic.

What to back up

Store	Importance	Recovery
Postgres	Critical	Required for any restore. Holds memory_entries, users, sessions, tokens, projects, audit log.
Qdrant	Optional	Re-index from warm-tier entries via /v1/admin/reindex (planned). Until then, snapshot and restore.
FalkorDB	Optional	Re-link from warm-tier on next dream cycle. Edges with kind = "vector_neighbour" will reform; manually-added edges (kind = "co_occurs") are lost without a backup.

In practice, back up all three. Postgres restoration without Qdrant + FalkorDB works (search degrades to keyword + vector with degraded: true for graph) but you want all three for full fidelity.

Postgres

pg_dump for full daily snapshots

PGPASSWORD=$POSTGRES_PASSWORD pg_dump \
  -h postgres -U novamem -d novamem \
  --format=custom --compress=9 \
  --file novamem-$(date +%F).dump

Restore:

PGPASSWORD=$POSTGRES_PASSWORD pg_restore \
  -h postgres -U novamem -d novamem --clean --if-exists \
  novamem-2026-05-05.dump

WAL archive for point-in-time recovery

If you need PITR (e.g. recover from accidental delete to 5 minutes before), enable continuous archiving:

# postgresql.conf
wal_level = replica
archive_mode = on
archive_command = 'aws s3 cp %p s3://my-bucket/wal/%f'

Combine with a base backup via pg_basebackup and you can restore to any LSN.

Qdrant

# Snapshot a collection
curl -X POST http://qdrant:6333/collections/novamem_acme_default/snapshots

# Download
curl http://qdrant:6333/collections/novamem_acme_default/snapshots/<name> > snap.tgz

# Restore
curl -X PUT http://qdrant:6333/collections/novamem_acme_default/snapshots/upload \
  -F snapshot=@snap.tgz

Snapshot every collection (one per tenant × project × namespace). If your tenant count is large, automate via the Qdrant snapshot API.

FalkorDB / Redis

Standard Redis persistence:

• AOF (append-only file) — appendonly yes in redis.conf. Replay-based recovery, lower data-loss risk.
• RDB (point-in-time dump) — save 900 1 style. Smaller files, larger possible data loss.

Both can be combined. Restore by replacing dump.rdb / appendonly.aof and restarting.

Disaster scenarios

Scenario	Recovery procedure
Postgres lost (no backup)	Total memory loss. Don't run without a backup strategy.
Postgres corrupt	Restore latest dump → restart novamem → cold-tier still has older vectors but new writes are coherent.
Qdrant lost	Restart with empty Qdrant → /v1/admin/reindex (planned) re-embeds every warm entry from content. Slow but lossless.
FalkorDB lost	Restart empty → next remember rebuilds edges via linkVectorNeighbors. Older entries lose graph-tier hits until they get re-linked.
Whole cluster lost	Restore Postgres dump → Qdrant snapshots → FalkorDB AOF → start novamem. Verify with GET /health.

Operational tip

Time-bound your backup tests. A backup that hasn't been restored is hypothetical. Schedule a quarterly drill: restore yesterday's dump into a sandbox cluster, run /health, run a few memory_search calls, confirm signal scores match expectations.