Key and Value Operations

Upstash Redis stores data as keys with strings, hashes, lists, and sets. A Beam agent reads and writes these records on demand: when a workflow needs a value, it fetches the key; when state changes, it writes the update under your approved rules. Writes that would overwrite protected keys, or reads returning unexpected types, get held for a person to confirm. The agent logs each operation so state stays predictable across services, and anything outside the rule, like a key it does not recognize, goes to a human instead of the agent guessing at intent.

Key Expiration Control

Upstash Redis lets each key carry a lifetime so stale data clears itself. A Beam agent reads current expiry settings and adjusts them according to your retention rules, expiring session tokens, caches, and temporary counters on schedule. When a key that policy says should persist is about to lapse, or an expiry would drop data another service still needs, the agent pauses and notifies an engineer. Approved changes are applied and recorded, keeping the store lean while a person decides the cases where automatic cleanup could remove something important to a running workflow.

Counter and Rate Limiting

Redis increments make it a fast home for counters, quotas, and rate limits. A Beam agent reads these tallies to check usage, then increments or resets them as your rules allow, for example capping API calls per customer or counting daily events. If a counter crosses a threshold you set, the agent notifies the right team; if a reset would erase figures needed for billing, it routes the case to a person first. Every change is written back and logged, so limits stay enforced without an engineer poking at keys by hand.