Skip to main content

A game or payment provider looks broken

Several players report the same game won't load, or cash outs through one payment company have all stopped. This guide is how to find out what's actually happening.

The screens are explained in Providers and their health.

Before you start​

  • Admin or super admin is fine. Both roles can press the test buttons here, so please read the warnings below even if you're new.
  • Have the provider's code, roughly when it started, what players are actually reporting, and the IDs of affected records.
  • Check the environment badge. Checks are read-only; test-account setup creates a real external account after super-admin confirmation.

Working through it​

  1. Open /admin/providers and select the game.

  2. Read before you click anything. The overall status, how recent the information is, what each individual test last said, the state of the test accounts, and any evidence shown. Very often you already have your answer here, without running anything.

  3. Read the same popup's enabled-action chips. Open Technical details only when you need diagnostic facts or test-account housekeeping.

  4. Check what's stuck. Look at /admin/provider-transfers and /admin/reconciliation for movements sitting in unknown or failed states. This tells you how many real players are affected, which is what everyone will ask you first.

  5. If you need to run a test, start with the gentlest one. Reads first. A game-launch check is also available when supported. Load, full-system and account-create tests no longer run.

    Read and launch checks move no money. Setting up a missing test account is a separate, confirmed super-admin action.

  6. Only run a test whose prerequisites are met. If the panel says something's missing, the test can't give a meaningful answer anyway. Wait for the result to come back on the row before pressing anything else.

  7. Name what you're seeing. Is it: information that's just old, a read problem, a game-launch problem, an account-creation problem, a load/redemption problem, a test account that didn't reset, or the provider being down entirely? Naming it well is most of the handover.

  8. Write it up. Test/run IDs, what each said, how fresh the data is, which transfers and games are affected, and who's picking it up next.

About the test accounts​

No test account exists yet. Set up creates one — and it's a real account at the outside company that keeps existing afterwards. If it times out, stop. We now don't know whether it was created. Write down the time and what you saw, and get it checked before trying again. Retrying can leave duplicate accounts on the provider's system.

A test account didn't reset properly (drift or restore failure). Quarantine it before it's used again, and keep the evidence. Quarantining takes it out of the automatic checks — it doesn't delete anything at the provider.

Acknowledge just silences an alert you've already looked at. It doesn't fix anything and it doesn't make a provider healthy. Don't use it to make a red row go away.

Stop and escalate when​

  • Real player transfers are sitting at unknown_requires_reconcile or failed_internal. Those are for investigation, never for retrying.
  • You're not certain which environment or external account you are looking at.
  • The capability list says one thing and the provider is clearly doing another.
  • The provider's own staff ask you for credentials, or for raw player or payment data. Don't send it. Route it through whoever owns the vendor relationship — this is exactly how a legitimate-sounding request turns into a real problem.

Two rules worth memorising​

A timeout is not a result. It isn't a success and it isn't a failure. It means we don't know, and the next step is evidence, not another click.

Never bulk-retry unknown movements. Retrying a batch of things that might already have gone through is the most expensive mistake available here. If dozens of records are stuck, that's a bigger conversation, not a bigger click.