Appearance
Go-live checklist
Production is a separate environment with live data. Most of what it needs is set up by Salaxy, and your code mainly switches hosts and uses a production certificate. Work through this list in order; each item links to the topic that explains it. Examples use the test environment; see Environments for demo and production.
Before production
Tick each item. Items 1 to 5 are one-time setup; ask Salaxy for the production side early, because Salaxy does the provisioning.
| # | What must be true | Who | Topic |
|---|---|---|---|
| 1 | Production partner account and technical account exist. They are separate from the test ones, with their own account ids | Salaxy, with you | Partner and technical accounts |
| 2 | Production certificate is created and its private key is in your backend secret store. A test certificate fails in production | You | Partner and technical accounts |
| 3 | Payment channel is provisioned in production and you have verified it | Salaxy, with you | Payment channels |
| 4 | Service models are provisioned in production, and you have agreed which customers get which model | Salaxy, with you | Service models |
| 5 | Accounting models are provisioned in production and linked from the service models | Salaxy | Accounting models |
| 6 | Hosts are switched from test to production, see below | You | Environments |
| 7 | Customer and accounting firm account creation works in production | You | Customer accounts, Accounting firm accounts |
| 8 | Smoke test passed, see below | You, with Salaxy | the topic pages you use |
| 9 | Customer care and escalation path is named, see below | You, with Salaxy | Developer FAQ |
The order of the topics you build is in Integration overview and Integration in a day with a coding agent. Items 1 to 5 are the same topics, with the production values: the test values are not reused.
Switch from test to production
Production is its own environment with its own data. Change these together, and keep the host settings in configuration, not in code.
| What | Test | Production |
|---|---|---|
| REST API base | https://test-secure.salaxy.com | https://secure.salaxy.com |
| Token endpoint | https://test-secure.salaxy.com/oauth2/token | https://secure.salaxy.com/oauth2/token |
| Embed host | https://test-embed.palkkaus.fi/ | https://embed.palkkaus.fi/ |
| Certificate | Test certificate | Production certificate, signed by the production root |
| Account ids (including the channel's owner account), service and accounting model ids | Test values | Production values from Salaxy |
- Do not reuse a test certificate or a test id in production. Production certificates are verified against a different root, and ids are different.
- The payment channel key does not change. It is fixed in code and is the same in test and production; only its owner account differs. See Payment channels.
- Demo does not exercise the production certificate path. Demo shares test data and uses the test root; only production proves the production certificate.
- Switch API and embed hosts together if you use the embedded white label.
- Authentication is unchanged: the same signed JWT flow, only the host, the certificate and the account ids differ. See REST authentication.
Smoke test
A suggested checklist, not a fixed Salaxy procedure. Run one complete case in production before real customers use it. Production sends real payments and real payslips, so use a case that is meant to be paid.
- Sign in with the production certificate and read something small.
- Create a customer account and an employee (Customer accounts, Employees and employments).
- Run the loop your integration uses: a payroll (Salary calculations), or the invoice and payment loop (Invoices and payments), through to the paid status (Payroll lifecycle).
- Read the result your system needs, for example accounting data or the salary slip.
- If you use the embedded white label, open it with the production embed host and check that sign-in works (URL parameters).
- Check that you log request ids and can quote them to Salaxy (Limits, logging and correlation).
Customer care and escalation
Two flows are easy to confuse. Name a person for each before go-live.
| Flow | Who asks | Who answers |
|---|---|---|
| End-customer care — a payroll question from a customer company or an accounting firm | The end customer | Usually you first, and you escalate to Salaxy what you cannot resolve. Your partner contract sets the actual split, and Salaxy publishes no standard response-time commitment |
| Escalation to Salaxy — a developer or operator needs something from Salaxy | You | Salaxy support |
Reasons to escalate to Salaxy, during development and after go-live:
- Missing test data or test accounts (Test endpoints and test data).
- Provisioning: technical account, payment channel, service and accounting models.
- Whether your payment channel has an outbound status trigger.
- A manual monthly-cycle run in test, where the cycle does not run end to end (Monthly cycle).
- A server-side log lookup: send the request id and an approximate time (Limits, logging and correlation).
Decide early whether Salaxy may be named to your end customers. In an embedded white label integration the customer sees your brand, so first-line staff need to know what they can say. Help channels are on Support.
Checklist
- Items 1 to 9 above are ticked, with production values and no test certificate or test id left.
- Hosts, certificate and ids come from configuration, one set per environment.
- A smoke test, for example the one suggested above, has run through the whole loop your integration uses, in production.
- You have a named person for end-customer care and a named Salaxy contact for escalation.
- You log request ids and know how to quote them to Salaxy.

