Durable Workflow Orchestration
Temporal Cloud pricing: how do Actions, Active Storage, and Retained Storage billing work?
7 min read
Most teams evaluate Temporal Cloud because they’re tired of over‑provisioning infrastructure “just in case.” Temporal Cloud flips that model: you pay for what your code actually does, not for idle servers or oversized databases. To understand your bill, you only need three primitives: Actions, Active Storage, and Retained Storage.
Quick Answer: Temporal Cloud pricing is consumption‑based. You pay for:
- Actions – the units of work processed by the Temporal Service.
- Active Storage – Workflow and system data for running and recently completed Workflows.
- Retained Storage – long‑term storage for older Workflow histories you choose to keep.
All three map directly to Temporal’s Durable Execution model, not to servers, clusters, or replicas you have to manage.
Frequently Asked Questions
What are Actions in Temporal Cloud pricing, and what counts as an Action?
Short Answer: An Action is a unit of work processed by the Temporal Service—things like starting a Workflow, scheduling or completing an Activity, firing a timer, or delivering a signal. Your Actions count reflects how much durable coordination the Service is doing for your code.
Expanded Explanation:
Temporal Workflows produce a stream of events: a Workflow starts, schedules Activities, waits on timers, receives signals, updates search attributes, and eventually completes or times out. Each of those state transitions is driven by the Temporal Service and recorded in the Workflow’s event history so it can be replayed and recovered after failures.
Temporal Cloud measures that work as Actions. Conceptually, whenever the Service:
- Accepts a new Workflow or Activity task,
- Records a state transition in Workflow history,
- Evaluates timers, schedules, or signals,
- Orchestrates retries, timeouts, or heartbeats,
it’s performing Actions on your behalf. The pricing model tracks how many such operations happen, not how many pods you run. This aligns cost with what you actually care about: “How much durable orchestration did my app need?”—not with CPU guesses or cluster size.
Key Takeaways:
- Actions are service‑side operations that drive Workflow execution and durability.
- Your Actions count scales with real work done (Workflows started, Activities run, timers fired), not with provisioned hardware.
How does Temporal Cloud bill for Actions in practice?
Short Answer: Temporal Cloud bills Actions on a consumption basis: the more Workflows, Activities, and timers you execute, the more Actions you use. Idle periods cost you almost nothing because you are not paying for pre‑allocated capacity.
Expanded Explanation:
In self‑hosted distributed systems, you size clusters for peak traffic, then pay for that capacity 24/7—whether you’re running 10 Workflows or 10 million. Temporal Cloud takes a different stance. The Service is operated, scaled, and replicated for you. You only pay for the work it performs, measured as Actions.
That means a small team can start with low volume and modest cost, then grow to hundreds of thousands of Actions per second without re‑architecting clusters or negotiating for bigger instances. Temporal Cloud can automatically scale to 300k+ Actions per second, and your bill simply reflects how much of that you actually used.
Steps:
- Model your process as code. Write Workflows and Activities using Temporal’s SDKs (Go, Java, TypeScript, Python, .NET).
- Run in Temporal Cloud. Your Workers stay in your environment; Temporal Cloud runs the Service and coordinates execution.
- Review usage. Monitor Actions in your account to see how Workflow volume, retries, and timers translate to consumption‑based costs.
What’s the difference between Actions, Active Storage, and Retained Storage?
Short Answer: Actions measure how much work the Service does; Active Storage covers state for live and recent Workflows; Retained Storage covers long‑term Workflow histories you keep for audits, debugging, or analytics.
Expanded Explanation:
Temporal is built around durable event histories. Every Workflow’s full execution is recorded so it can be replayed after crashes and inspected later in the Web UI. Storage pricing separates that lifecycle into two phases:
- Active Storage is for Workflow state and history that’s “hot” and used to power running Workflows, recent completions, and operational queries.
- Retained Storage is for “cold” histories you preserve for months or years after Workflows are done, usually for compliance, forensic debugging, or analytics.
Together with Actions, these three dimensions map cleanly to Temporal’s internals:
- Actions → coordination work: scheduling, retries, timers, signals, completions.
- Active Storage → data needed to keep Workflows live and debuggable right now.
- Retained Storage → older histories you choose not to delete.
Comparison Snapshot:
- Option A: Actions – Pay for actual orchestration work done by the Service.
- Option B: Storage (Active + Retained) – Pay for how long and how much Workflow history you keep.
- Best for: Teams that want cost to track business traffic and data retention needs, not hardware sizing.
How does Active Storage billing work in Temporal Cloud?
Short Answer: Active Storage billing is based on how much Workflow and system data you keep “hot” in Temporal Cloud for running and recently completed Workflows.
Expanded Explanation:
Every Temporal Workflow has an event history: a sequence of Actions that describes its full life—from start to completion. To recover from failures, the Service must keep that history durably stored and quickly accessible. That’s what Active Storage represents.
While a Workflow is running—or recently finished and still within the active retention window—its event history, search attributes, and related metadata reside in Active Storage. Temporal Cloud replicates this data and optimizes it for low‑latency reads and writes, so task dispatch, timers, and replay are fast and predictable.
You’re billed on the volume of this active data. The more Workflows you have in flight (and the more events per Workflow), the more Active Storage you use. When Workflows age out of your active retention period (or you explicitly delete them), their histories can move out of Active Storage—either to Retained Storage (if you keep them) or be removed entirely.
What You Need:
- A rough sense of concurrent Workflows and how event‑heavy they are (e.g., a few steps versus thousands).
- A retention policy for how long you typically need immediate, low‑latency access to recently completed Workflows in the Web UI and APIs.
How does Retained Storage billing work, and why keep old Workflow histories?
Short Answer: Retained Storage billing is based on how much historical Workflow data you choose to keep long term after it’s no longer active. It exists so you can retain full execution histories for audits, debugging, or analytics without overpaying for “hot” storage.
Expanded Explanation:
Once Workflows complete and are no longer needed for daily operations, many teams still want to keep their histories:
- To debug rare or customer‑specific issues months later by replaying the exact sequence of steps.
- To satisfy regulatory or internal audit requirements for money movement, orders, or compliance flows.
- To analyze execution patterns across millions of runs for optimization or capacity planning.
Temporal Cloud exposes this as Retained Storage. It separates long‑term archival from the high‑performance storage needed for live Workflows, so you’re not paying premium rates for data that’s rarely accessed. You control how long histories are retained and at what granularity.
Because Temporal workflows are inherently auditable—every state change is an event—long‑term retention provides a far richer forensic trail than logs alone. Instead of stitching together partial log lines and approximate timestamps, you can inspect or replay the exact Workflow in the Temporal Web UI.
Why It Matters:
- Auditability and compliance: Keep end‑to‑end histories of money movement, order lifecycles, and high‑risk operations.
- Deeper debugging and analytics: Use full, replayable histories instead of hunting through logs and dashboards.
Quick Recap
Temporal Cloud pricing is built around the actual primitives of Durable Execution, not guesswork about hardware. Actions capture how much orchestration work the Service does. Active Storage covers the hot data that keeps your Workflows running and debuggable in real time. Retained Storage lets you keep complete Workflow histories as long as you need them, at a cost optimized for long‑term retention. You run your Workers in your own environment, Temporal Cloud runs the Service in 11+ regions with built‑in replication and disaster recovery—and either way, Temporal never sees your code.