Reserved meters
Six electricity meters, one outcome each.
In test mode these are the only meters that work. Anything else is refused with "This is not one of the sandbox test meters."
| Meter | Verify | Order | Ends as |
|---|---|---|---|
11111111111 | passes | 202 | delivered, with a token |
22222222222 | passes | 422, order in data | refunded |
33333333333 | passes | 422, data is null | nothing is created |
44444444444 | passes | 202 | processing, then delivered after 15s |
55555555555 | passes | 202 | processing, then refunded after 15s |
66666666666 | fails | not orderable | n/a |
Use any disco_code you like, so long as it is not empty, and any amount
above zero.
What each one is for
11111111111, delivered. The happy path, and the one to build against
first. Settles immediately, so you get a token in the order response.
22222222222, refunded inline. The order fails on arrival. You get a 422
whose data holds a settled order with state: "refunded". There is no webhook,
because the order was already finished when you were told about it. This is the
shape most integrations get wrong, so test it early.
Important
The live path produces the same inline-failure shape with state: "failed"
rather than refunded. The two modes genuinely differ here. Handle both
terminal states rather than matching on one.
33333333333, refused before anything is charged. A 422 with data: null. No order exists, nothing was charged, and your reference is still
free, so you can reuse it. The only reserved number that behaves this way.
44444444444, delayed delivery. Comes back processing and settles 15
seconds later. The one to test your polling and your webhook handler against,
because it is the only reserved meter that forces you to wait. If your code
calls once and treats processing as done, this is the meter that proves it.
55555555555, delayed refund. processing, then refunded 15 seconds
later, with a webhook carrying refund_amount. Test what your customer sees
between "we are buying it" and "you got your money back".
66666666666, verify fails. Meter verify answers 422. Use it to check
what your UI does when a customer types a meter number that does not exist. It
cannot be ordered, because without a successful verify you have no
access_code.
The access code is checked
Sandbox meter verify returns an access_code like
sbx_4f1c9a2b7e6d3085a1b4c7d2, deterministic for that meter, and the order
checks it. You cannot shortcut the verify step in test mode any more than in
live mode.
{
"name": "Sandbox Customer",
"address": "1 Test Close, Sandbox",
"phone_number": "08000000000",
"meter_number": "11111111111",
"short_code": "IKEDC",
"meter_type": "PREPAID",
"access_code": "sbx_4f1c9a2b7e6d3085a1b4c7d2"
}The customer is always the same, and short_code and meter_type come back in
upper case whatever you sent.
Tokens and units
A delivered sandbox order gets a token of the form TEST-4821 9033 7715 2064,
worked out from your partner id and your reference, so the same reference gives
the same token every time.
token_units is amount / 70, rounded to one decimal place. 5,000 naira gives
71.4 units. The 70 is a fixed sandbox figure, not the live tariff.