Business observability is software that follows each business transaction, such as a payment, across all the systems it passes through. It shows where transactions are failing, counted in business steps rather than in servers. It is also called business journey monitoring, or business transaction monitoring. In India, its best-known users are banks and payment companies.
Why is it needed? An ordinary observability platform shows technical data: a slow service, a full disk, an error in a log. But a bank's managers ask different questions. How many UPI payments failed in the last hour? At which step did they fail? Did they fail for customers of one partner bank only? How much money is stuck? To answer these questions, someone has to join the technical data from a dozen systems into one view of each payment.
Think of the steps in one UPI payment. A customer pays a shop from her UPI app. The request enters her bank through the bank's UPI switch. The switch checks her UPI PIN. It asks the core banking system to debit her account. Then it replies to NPCI, which runs UPI, and NPCI asks the shop's bank to credit the shop. Each of these systems writes its own records, in its own format, and with its own number for the payment. Business observability reads all of them. It joins the records of each payment, using the payment's transaction reference number, and draws the payment's journey step by step.
Here is a typical case. A mid-sized bank in Kolkata handles about 40 lakh UPI payments a day. At 7:15 p.m., its dashboard shows that the share of successful payments has dropped from 99.2% to 95.8%. The tool breaks down the failures by step. Almost all of them happen at one step: "debit request sent to the core, no reply within 10 seconds". It breaks them down further, by server. They all come from one of the bank's two core banking servers. The operations team moves the traffic away from that server, and by 7:40 p.m. the success rate is back above 99%. The tool also lists the 9,300 payments in which the customer's account was debited but the shop was never credited. These customers must get their money back. The RBI sets a time limit for such reversals, and the bank must pay the customer compensation if it misses the limit.
Why do banks watch this so closely? NPCI tracks how many UPI payments each bank fails for technical reasons, and the figures are made public. A bank with a high failure rate looks unreliable to its customers and to the regulator.
Business observability uses the same telemetry as other monitoring tools. What it adds is the business meaning. It reports "debited but not credited", instead of a technical error code such as "HTTP 504". The same idea works for loan applications, insurance claims or train bookings: anything that passes through several systems in fixed steps.