Skip to main content
On xpander’s cloud, xpander operates the data tier. On Hybrid and Air-Gapped installs it runs in your cluster, and two of the stores can be replaced by managed services you already run.

What each store holds

In-cluster is the default: chart-managed StatefulSets on your StorageClass, SSD-backed. Production sizing on Air-Gapped: xpander’s PostgreSQL 50 GiB, Supabase PostgreSQL 50 GiB, Supabase Storage 50 GiB and up, MongoDB 20 GiB, Redis 8 GiB, plus about 50 GiB for the registry mirror. Set --storage-class <name> if the cluster default is not the one you want.

What is stored where

Supabase’s PostgreSQL (database postgres)

xpander’s PostgreSQL 16 (database xpander)

Isolation between agents is a schema per agent in this one database.

MongoDB

Looking inside from the cluster

The passwords come from the xpander-generated-keys secret. Nothing below prints one.
At the mongosh prompt:

Bringing your own PostgreSQL and Redis

Installer flags or values, with the passwords passed as --set so they never land in Git:
  • PostgreSQL: create the role and database before the install. Keep CREATEDB on the role. Agent data lands as one schema per agent (ag_<agent id>) inside the xpander database. keygen.overrides.POSTGRES_PASSWORD is how xpander learns the password you set.
  • Valkey or MemoryDB: the endpoint must ship the RediSearch commands, or knowledge bases accept writes and return nothing. On MemoryDB that is Valkey 7.3 on the parameter group with the search module.
  • Probes: before installing, the installer launches a probe pod inside your namespace and runs pg_isready and redis-cli ping from where xpander will connect. A failed probe is a decision point, not an abort.
Details and the reference customer’s RDS and MemoryDB sizes are on Air-gapped: data tier.

MongoDB

MongoDB has no external mode. It holds the skill catalog’s OpenAPI operations and the monitoring and audit stores, and runs in-cluster on every self-hosted deployment. Its volume is part of the chart’s footprint; back it up with your cluster’s volume snapshots.