Temporal
Temporal is an open-source durable execution platform that lets developers write business logic as ordinary code while the server durably persists workflow state, replays execution history, and automatically retries failed activities. Multi-language SDKs and a self-hostable server mean long-running or fault-sensitive processes need no custom checkpointing logic.
What gets deployed with it
- PostgreSQL HA Cluster — required Temporal requires a durable relational persistence backend for workflow state, task queues, and history records. PostgreSQL is the recommended production backend.
Baseline topologies
| Topology | Minimum nodes | Roles |
|---|---|---|
| pooled-cluster | 3 | temporal-cluster |
| per-service-scaled | 4 | temporal-frontend, temporal-history, temporal-matching, temporal-worker |
What drives the cost
- Peak concurrent workflow executions
- Peak number of workflow executions running simultaneously across the cluster. Drives History service memory and task dispatch throughput.
- Active scheduled workflows
- Total number of workflows with active cron or interval schedules. Drives Matching service and scheduler heartbeat load.
- Workflow history retention (days)
- How long completed workflow execution history is retained in the persistence backend. Affects database storage footprint.
Where it runs
bare_metal, dedicated, cloud_vm
Size Temporal and get a cost estimate