Temporal on Railway
Self-hosted Temporal server for durable workflows written as code, with Postgres and a Web UI, on Railway.
Temporal is an open-source durable execution platform. You write workflows as ordinary code in Go, Java, TypeScript, Python, .NET, PHP or Ruby, and Temporal records their progress so they resume after crashes, retries and long waits without losing state.
Teams use it for payment and order flows, long-running processes with timers and approvals, background jobs, data pipelines and AI agent orchestration. It is often compared with queue-based job systems and with workflow engines such as Airflow; the difference is that workflow logic lives in your own code.
The Temporal server is MIT licensed. Temporal Cloud is the vendor's hosted option.
What Temporal can do
- Workflows as code in several languages
- Automatic retries and durable timers
- Full workflow history
- Web UI for inspecting workflows
- Namespaces for separation
- SQL or Elasticsearch visibility
What people use it for
- Order, payment and subscription flows that must finish despite failures
- Long-running processes with timers and human approvals
- Reliable background jobs and data pipelines
- Orchestrating AI agents and LLM tool calls
Videos
Running Temporal on Railway
Two templates are listed. The Tracklayer template runs Temporal Server 1.32.1 (all roles in one service) with PostgreSQL for persistence and visibility, no Elasticsearch, a job that registers the default namespace, and the Web UI behind a Caddy gateway with basic auth. A community template adds Elasticsearch and an S3 history archive. Workers are not included; deploy them in the same project and connect over the private network. Cost follows the server, Postgres and, if used, Elasticsearch.
Temporal templates on Railway
Tracklayer templates are listed first. Other Temporal templates follow, ordered by total deployments on Railway. How this works.
Deployment numbers as of 2026-11-20.
Questions about Temporal
Which Temporal template should I start with?
The Tracklayer template is the simpler one: Postgres only, with a password-protected Web UI. The community template adds Elasticsearch and S3 archiving, but its listing gives few details, so review its variables first.
How do my workers connect?
Deploy them in the same Railway project and set TEMPORAL_ADDRESS to the temporal service's internal address on port 7233 and TEMPORAL_NAMESPACE to default.
Can I reach the server from outside Railway?
Only by adding a TCP proxy on port 7233. The frontend has no authentication by default, so do this only together with mTLS.
What limitation should I know before going live?
The history shard count is fixed at first start and cannot be changed later, so decide it before real use.
How do I upgrade?
Change the server and admin-tools tags in the service Dockerfile, one minor version at a time. The pre-deploy step applies schema changes first.
Guides
Tracklayer maintains the templates marked as such; Temporal is a trademark of its owners and this site is not affiliated with them.