Provider health monitoring is the continuous measurement of whether each payment provider a business relies on is up, fast and approving transactions at its normal rate. It tracks authorization rate against the same hour in previous weeks, 95th and 99th percentile latency, technical errors, decline code shifts, webhook delays and 3DS success, often split by country, card brand, BIN range and method. Passive monitoring reads live transactions, while synthetic test calls cover quiet hours. When a provider drifts from its baseline, alerts fire or circuit breakers and smart routing move traffic elsewhere. Merchant-side data catches partial degradation that public status pages miss, supports SLA claims, and helps EU financial entities meet DORA monitoring and incident-reporting requirements.
What is payment provider health monitoring measuring
Availability alone is a weak signal, because a gateway can answer every request and still decline half of them. Useful monitoring tracks several metrics per provider, often split by country, card brand, BIN range and payment method:
- Authorization rate, compared with the same hour of previous weeks.
- Latency at the 95th and 99th percentile, not only the average.
- Technical error rates: HTTP 5xx responses, timeouts, connection failures.
- Share of specific decline codes, such as a sudden rise in "issuer unavailable".
- Webhook delay between authorization and notification.
- 3DS success rate and the share of challenges abandoned.
How does provider monitoring work in practice
Two data sources feed it. Passive monitoring reads every real transaction as it completes, which catches problems that only affect certain issuers or currencies. Active or synthetic monitoring sends scheduled test calls to each endpoint, which covers quiet periods such as 4 a.m. when there is too little live traffic to measure.
Example: a merchant's approval rate on PSP A normally sits at 88% for German Visa debit cards. Within ten minutes it drops to 71% while PSP B holds steady at 87%. The monitor flags the gap, and the smart routing engine moves that segment to PSP B until PSP A recovers.
Payment gateway uptime monitoring vs status pages
Providers' public status pages are updated manually and often late, and they report platform-wide outages rather than a failing acquirer connection in one market. Payment gateway uptime monitoring built on the merchant's own traffic sees partial degradation first. Evidence gathered this way also supports claims under the provider's service level agreement.
Turning signals into action
Monitoring feeds automated responses. A circuit breaker stops sending traffic to a provider after a set failure threshold and periodically tests whether it has recovered. Cascading retries a soft-declined payment on a second provider. Timeout strategy settings define how long to wait before treating a call as failed, which matters because a timed-out authorization may still have succeeded at the issuer.
Regulatory angle
For EU financial entities, DORA requires monitoring of ICT third-party providers and classification and reporting of major ICT-related incidents. A payment institution that can show per-provider health data, incident timelines and failover tests is in a much stronger position with its regulator, and with a buyer assessing operational resilience.
See also: Circuit Breaker Pattern.