A staging environment next to production
On Livops, Add environment on the project page makes a staging copy of your production services, each with its own address, its own variables and its own empty database, so a test never touches production data. Promote to production then ships the exact release you tested, without rebuilding. A staging environment costs half; pull request previews are billed like regular services.
By Livops team, the people who build and run LivopsUpdated
A staging environment is a second copy of your app where a new version runs against its own data before customers see it. This page explains what a staging environment needs to be useful, shows how to add one next to production on Livops and promote a tested release, and compares how Render, Railway and Vercel do the same thing, from their own docs.
What makes a staging environment useful?
Its own data
Staging that writes to the production database is not staging. It needs its own database, empty or seeded, so a migration that goes wrong breaks nothing real.
Its own variables
A payment key in test mode, a different bot token, a webhook address that goes nowhere: staging needs the same variable names as production with different values, and no way to read production’s.
The same build that goes to production
If production builds the code again after you have tested it, what reaches customers is a new build: a dependency can resolve differently, or a build-time variable can change it. The strongest guarantee is to ship the files you tested.
Its own address
Testers and webhooks need a URL that does not change between versions.
Add a staging environment on Livops
Every project starts with a production environment. On the project page, click Add environment and name it — usually staging. The new environment:
- gets copies of the services in production, each with its own URL;
- gets its own empty database for each database type production has — no data is copied, and staging never writes to the production database;
- copies the variables; any that cannot be copied, such as one pointing to another database, are shown as not set for you to fill in.
The database’s address is written into the staging services by itself, as in production: staging’s DATABASE_URL points at staging’s database.
Which branch staging deploys
A staging service starts as a copy of the production one. The branch a service deploys from is in its Settings, as for any service connected from GitHub, and every push to that branch deploys it. For a staging that runs ahead of production, point it at a branch such as staging, test there, and ship to production with Promote.
Variables per environment
- Each service has its own Variables. Mark keys as secret and their values are never shown again, in the cabinet or the API.
- Values every service in one environment needs, such as
SENTRY_DSN, go in that environment’s shared variables; a service’s own value with the same name wins. - A value can reference another service’s variable or address, or a shared variable — see References. References resolve only inside their own environment: a service in
stagingcannot see one inproduction, so staging’s API address points at staging’s API.
Promote to production
When the version in staging is the one you want, press Promote to production.
- The running release is placed on the production service without rebuilding: the exact files you tested ship.
- Production keeps its own variables. Staging’s
DATABASE_URLdoes not carry over. - Static sites are the exception. Their
VITE_*andNEXT_PUBLIC_*variables are baked into the files at build time, so the site is rebuilt from the same commit with production’s variables.
The same from a script, with an API token:
curl -X POST https://livops.uz/api/projects/<staging-service-id>/promote \
-H 'Authorization: Bearer lvp_xxx'
The assistant can propose a promote as a card you press, and MCP clients have promote_release.
Pull request previews
When preview is on in the project settings, a separate environment opens for every pull request on GitHub and is deleted when the pull request closes. Preview environments have no analytics counters, so they do not skew your visitor numbers. A preview is billed like a regular service.
What it costs
Staging and other environments cost half of what the same services cost in production. A bot is 19 credits a month and an app 39 credits in production, and a static site is free. Credits are bought in so’m through Payme or Click, with no card; current prices are always in the cabinet.
Deleting an environment deletes its services and databases too, and can only be done in the cabinet, not with an API token.
How Render, Railway and Vercel do it
| Platform | Long-lived staging | Per pull request | What staging costs |
|---|---|---|---|
| Livops | Add environment: copies of the services, own variables, own empty databases | Preview environments, deleted when the PR closes | Half the production price; a preview like a regular service |
| Render | Environments inside a project; Hobby workspaces get up to two per project[1] | Preview environments on Pro or higher, defined in a Blueprint[2] | Services at the usual rates[2]; Pro is $25 a month plus compute[3] |
| Railway | Persistent environments, created as a duplicate or empty[4] | PR environments, deleted when the PR is merged or closed[4] | Billed by usage[5] |
| Vercel | Custom environments on Pro and Enterprise, or a preview branch on any plan[6] | A preview deployment for every branch push and pull request[6] | Pro includes 1 custom environment per project; more are $50 a month per five[6] |
Other providers’ details are from each provider’s own page, read on 20 November 2026; the dates are in Sources below.
Render
A Render project holds one or more environments, so staging and production services stay apart, each with environment variables and secret files only its own services can read[1]. An environment can be marked protected, so only admins make destructive changes, and its private network can be closed at its boundary[1]. Preview environments copy production for each pull request, need a Pro workspace and a Blueprint file, and start their datastores empty[2]; they are billed like other services, by the second[2]. Render’s protected environments are something Livops does not have.
Railway
Every Railway project starts with a production environment, and changes to a service are scoped to one environment[4]. A persistent staging can auto-deploy from a staging branch with its own variables[4]. A duplicated environment copies services, variables and configuration, and its services wait for you to approve the staged changes before deploying[4]. PR environments are switched on under Project Settings → Environments[4]. Usage is billed by the second[5].
Vercel
Vercel has Local, Preview and Production environments by default, and each environment can have its own variables[6]. Every push to a branch other than the production branch, and every pull request, gets a preview deployment with a branch URL and a commit URL[6]. A named staging with branch tracking and its own domain is a custom environment, on Pro and Enterprise[6]; Pro is $20 a month[7]. Vercel can also stage a production build and promote it without rebuilding[6], as Livops’s promote does; promoting a preview deployment, though, rebuilds it with production’s variables[6]. Its previews are the most automatic of the four for frontends: no setting to turn on.
Staging on Livops
- On the project page, click Add environment and name it
staging. - Open Variables in the staging services and fill in any marked not set; give test keys their test values.
- In each staging service’s Settings, check the branch it deploys from, and push to it.
- Load test data into staging’s own database: open its Connect from your computer, start the SSH tunnel, and run your seed script or
pg_restorethrough it. - When the release is right, press Promote to production.
- Optional: turn on preview in the project settings for an environment per pull request.
See Environments and promote, Variables and Pricing.
Questions
Does staging share the production database?
No. When you add an environment, it gets its own empty database for each database type production has. No data is copied, and staging never writes to the production database.
Are variables copied to staging?
Yes, when the environment is created. Variables that cannot be copied, such as ones pointing to another database, are shown as not set and you fill them in. After that, each environment keeps its own values, and a reference to another service only sees services in its own environment.
Does promoting rebuild my app?
No. Promote to production places the release running in staging on the production service as it is, so the exact files you tested ship. Production keeps its own variables. Static sites are the exception: their variables are baked in at build time, so the site is rebuilt from the same commit with production's variables.
How much does a staging environment cost?
Half. Staging and other environments cost half of what the same services cost in production, paid monthly in credits. A pull request preview is billed like a regular service. Current prices are always in the cabinet.
Can I have a preview for every pull request?
Yes. Turn on preview in the project settings, and a separate environment opens for each pull request on GitHub and is deleted when the pull request closes.
Can I delete a staging environment with an API token?
No. Deleting an environment also deletes its services and databases, and an API token cannot do it. Delete it in the cabinet.
Sources
- Render — Projects and environments (render.com). Checked .
- Render — Preview environments (render.com). Checked .
- Render — Pricing (render.com). Checked .
- Railway — Environments (docs.railway.com). Checked .
- Railway — Pricing (railway.com). Checked .
- Vercel — Environments (vercel.com). Checked .
- Vercel — Pricing (vercel.com). Checked .
Put your project online today
Connect GitHub or drop a folder. No card is asked for.
Add a staging environmentRelated
- Deploy from GitHub on every pushConnect a repository through the Livops GitHub App, pick a branch, and every git push to that branch builds and ships a new version, with the build log in the cabinet. A failed build leaves the running version alone, and a previous version is one click away in the Deploys list. To ship only after your tests pass, turn on Deploy after CI passes or call a deploy hook from GitHub Actions.
- Managed PostgreSQL hosting in 2026: the options comparedA production PostgreSQL needs backups you can restore, a connection limit your app stays under, the app in the same region, and a supported version. On Livops a PostgreSQL 15, 16 or 17 database runs inside your project, its DATABASE_URL is written into your services, and it is backed up every night; 500 MB is free, one per team, and 1 GB is 9 credits a month. Here is how Neon, Supabase, Render, Railway and a VPS compare.
- ECONNREFUSED / Connection refused: why the app cannot reach its database after deployNothing was listening at the address your app tried. When the address is 127.0.0.1 or localhost, the app never read the database's real address: DATABASE_URL is not set, is not read, or is read during a build that has no database. Attach the database, read DATABASE_URL (or REDIS_URL, MYSQL_URL) at runtime, run migrations before the start rather than in the build, and keep the connection pool under the server's limit.
- Livops vs Render: which one to deploy on in 2026Pick Render if your users are in the US, Europe or Singapore, you pay by card, and you want free web services to start with, autoscaling and larger managed Postgres. Pick Livops if your users are in Uzbekistan and you want to pay through Payme or Click, with no card, at 19 credits a month for a bot or 39 for an app.
- Livops vs Railway: which one to deploy on in 2026Pick Railway if your users are in the US, Europe or Southeast Asia, you pay by card, and you want usage-based billing with high-availability databases. Pick Livops if your users are in Uzbekistan and you want to pay through Payme or Click, with no card, at a fixed 19 credits a month for a bot or 39 for an app.
- Livops vs Vercel: which one to deploy on in 2026Pick Vercel for a Next.js or frontend app served to users around the world from its global network, paid by card. Pick Livops if your users are in Uzbekistan, you also run a Telegram bot or a backend that must stay up, and you want to pay through Payme or Click, with no card; static sites are free on both.