I would ask if after the customers are done with the payment, whether Swiggy receives the payment even though paytm doesn't give any confirmation.
If this is true, then there will be discrepancy between status of payments and money received. I mean, even though the payment shows as pending or failed, the money on Swiggy's account will be higher indicating a failure of confirmation of a payment gateway. In this case, Swiggy will know of the problem. It will examine all the payment gateways immediately and notice which of the payment gateways has failed and thus make alerts quicker.
I did like the candidates thoughts as he detailed the whole ordering procedure to arrive at the problem.
My initial question was, why didn't the interviewer give the problem right away?
My assumption: In real world, all one notices is fall in no. of orders and not whether payments have failed, or pending. If the orders fell by 0.1% they wouldn't have noticed it as it could have been attributed to randomness in periodicity. But suddenly, when the orders fall by 10% one day, the person has to find out the cause. So, there is no one to actually give the real cause.
But my question still remains, a dashboard of metrics that tracks customer time spent on the app, orders placed in the basket, orders not converted, one of the metrics will give the problem away as soon as it happens. Anyone would like to answer?
I would ask if after the customers are done with the payment, whether Swiggy receives the payment even though paytm doesn't give any confirmation.
If this is true, then there will be discrepancy between status of payments and money received. I mean, even though the payment shows as pending or failed, the money on Swiggy's account will be higher indicating a failure of confirmation of a payment gateway. In this case, Swiggy will know of the problem. It will examine all the payment gateways immediately and notice which of the payment gateways has failed and thus make alerts quicker.
I did like the candidates thoughts as he detailed the whole ordering procedure to arrive at the problem.
My initial question was, why didn't the interviewer give the problem right away?
My assumption: In real world, all one notices is fall in no. of orders and not whether payments have failed, or pending. If the orders fell by 0.1% they wouldn't have noticed it as it could have been attributed to randomness in periodicity. But suddenly, when the orders fall by 10% one day, the person has to find out the cause. So, there is no one to actually give the real cause.
But my question still remains, a dashboard of metrics that tracks customer time spent on the app, orders placed in the basket, orders not converted, one of the metrics will give the problem away as soon as it happens. Anyone would like to answer?