Queues. Streams. Sessions. Counters. Data with no upstream source to rebuild from — if you’re running AOF because losing it would be an incident, you’re already using your cache as a source of truth. Swytch Cloud makes that legitimate: durable storage that survives full-cluster restart, with point-in-time recovery. Not to keep the cache warm — to keep the truth safe.
Rebuilding cached values after a restart is what caches are for — warming a cache from durable storage would hand you yesterday’s answers while the database moved on. That is not what this is.
This is for the data that exists nowhere else. Your data plane keeps serving from your infrastructure; Swytch Cloud holds the durable log behind it — zero-knowledge, storing bytes it cannot read. Bring the whole cluster down and back, and the queues, streams, sessions, and counters return. Rewind to a point in time when you need yesterday back on purpose.
# data with no upstream
r.xadd("orders", {"id": 8812}) # event log
r.lpush("jobs", payload) # work queue
r.hset("session:42", ...) # login state
r.incr("invoice:seq") # a counter that must not reset
# a restart must not erase theseIf losing it is an incident, it was never a cache.
For the data that exists nowhere else.
Bring every node down and back. Queues, streams, sessions, and counters return — not from a lucky replica, from the durable log.
A bad deploy or a fat-fingered FLUSHALL is not forever. Rewind to when the data was right.
Cloud holds the durable log without being able to read it. Your data plane serves from your infrastructure; the truth stays yours even while we keep it safe.
Sign up, connect your cluster to the control plane, and your queues, streams, and sessions stop being one restart away from gone. €20 starting credit on sign-up.
Using Redis as a source of truth today? Drop your address and we’ll send the durability write-up.