What Redis is for
Redis is optional, shared application infrastructure. It backs features that need a background queue or scheduled work rather than an immediate request/response — currently:
- Automatic CTR/FTR report generation. The daily, weekly, or monthly sweep that drafts Currency Transaction and Funds Transfer reports runs as a background worker. See Reporting.
- Notification emails. Every notification email — including the "email me when I need to take action" preference on My Account — is queued and delivered by the same worker. See Notifications.
Redis is not required for transaction monitoring decisions, KYC verification execution, interactive report preparation, or goAML XML generation. Those remain synchronous and unaffected if Redis is never configured. Nothing else in OneView is blocked by a missing Redis connection; the two features above simply stay unavailable until it is configured.
OneView does not include or package Redis, the same way it does not include PostgreSQL. Provision a Redis instance the client controls and connect OneView to it. The generated installation does not bundle a Redis container and does not publish a Redis port.
Configure the connection
Open Settings → Redis. redis:read permits viewing whether Redis is configured. redis:manage permits testing
and saving the connection.
- Enter the hostname and port. Username and password are optional, for a Redis instance that requires authentication.
- Enable TLS when the Redis server requires an encrypted connection.
- Select Test connection. OneView probes the connection without saving it and reports the connected Redis server version on success.
- Select Use this connection to save it. OneView re-verifies the connection at save time in case it drifted since the test, so a save can still fail even after a successful test.
Once configured, the page shows Redis is configured instead of the form. A user with redis:manage can select
Reconfigure to point OneView at a different Redis instance; there is no separate delete or disconnect action.
Data boundary
Redis holds only transient, non-durable job data — never PostgreSQL's authoritative records. A queued job carries just a reference (an ID and a job type); the worker re-reads whatever it actually needs from PostgreSQL when it runs a job. Job payloads never contain customer data, KYC data, report content, transaction detail, or credentials, and Redis is never a system of record. Losing Redis's contents does not lose data — it only delays or drops queued work, which the relevant feature (automatic report generation, or a notification email) can recover from once Redis is reachable again.
The Redis connection's own credentials, once saved, are not displayed again. Keep the approved value in the client's secret-management process for recovery and credential rotation, the same as PostgreSQL.
Client responsibilities
Provision, secure, back up, and monitor Redis the same way as any other piece of client-managed infrastructure. Do not expose Redis publicly. Because Redis here is non-durable working storage rather than a system of record, it does not need the same backup and point-in-time-recovery treatment as PostgreSQL — see Backups & recovery for what does require it.
Dependent features fail closed
If Redis is not configured, OneView keeps the dependent feature visibly unavailable rather than silently degrading:
- The SMTP email page (Settings → Email) will not accept a connection until Redis is configured, and links directly to Settings → Redis to resolve it. See Notifications.
- Enabling an automatic CTR or FTR generation schedule is rejected until Redis is configured, and Reporting Settings links to Settings → Redis to resolve it.
Configure Redis first if you plan to use either feature.