Platform status

Current status

Each row reflects the live health of a production API surface, derived from synthetic probes that run a real request against every region once a minute. Sandbox shares the same control plane but is reported separately where it diverges.

Trailing 90-day uptime

Availability is measured as the share of one-minute synthetic probes that returned a healthy result over the trailing 90 days, per service. Our published target is 99.95% for core money-movement surfaces (Apply, Decisioning, Payments) and 99.9% for the rest. A probe counts as failed only on a 5xx, a timeout above the service SLO, or a correctness check failing — client-side 4xx responses are excluded.

Decisioning and Identity show higher tail latency because both call third parties — Companies House, bureau data, the bank-data aggregator and the sanctions/PEP screening provider — inside the request. Where an upstream is slow, the platform degrades gracefully: a decision may be returned as referred with a data_incomplete reason rather than blocking, and the webhook follows once the upstream responds.

Status API

Everything on this page is also a machine-readable feed. Poll the summary endpoint for a uptime/incident snapshot, or subscribe a webhook to receive status.incident.opened, status.incident.updated and status.incident.resolved events so you can surface platform health inside your own product.

# No auth required — the status feed is public.
curl https://status.credicorp.co.uk/api/v1/summary

{
  "object": "status_summary",
  "updated_at": "2026-06-29T14:02:11Z",
  "overall": "degraded",
  "services": [
    { "key": "apply",       "state": "operational", "uptime_90d": 0.9999 },
    { "key": "decisioning", "state": "operational", "uptime_90d": 0.9997 },
    { "key": "payments",    "state": "degraded",    "uptime_90d": 0.9994 },
    { "key": "identity",    "state": "operational", "uptime_90d": 0.9998 },
    { "key": "accounts",    "state": "operational", "uptime_90d": 1.0000 },
    { "key": "webhooks",    "state": "operational", "uptime_90d": 0.9999 }
  ],
  "active_incidents": [
    { "id": "INC-2026-0188", "impact": "minor", "service": "payments" }
  ]
}

Recent incidents

A rolling log of incidents over the last 90 days, newest first. Each carries an impact level (major minor resolved), the affected service, a timeline, and the root-cause summary once the post-incident review is complete.

INC-2026-0188 — Elevated PISP disbursement latency

Minor   Payments   ·   Investigating · opened 29 Jun 2026, 13:41 UTC

• 13:41 — Synthetic probes detected p99 disbursement-initiation latency above SLO on the Lloyds PISP gateway. Status set to degraded.
• 13:58 — Confirmed isolated to one provider. Affected disbursements automatically failing over to an alternate rail; no funds at risk and no failed payments.
• 14:02 — Monitoring. Borrowers may see disbursement land a few minutes later than usual. The loan.funded webhook fires as normal on completion.

INC-2026-0181 — Webhook delivery delays in eu-west-2

Resolved   Webhooks   ·   12 Jun 2026 · duration 38 min

• Impact — A backed-up delivery queue delayed outbound webhooks (e.g. decision.completed, loan.funded) by up to 11 minutes. No events were lost; all were delivered, late, and signatures verified normally.
• Root cause — A deploy reduced delivery-worker concurrency. The backlog drained once concurrency was restored and an extra worker pool was added.
• Follow-up — Queue depth now alerts at a lower threshold; worker concurrency is pinned outside the deploy config. If you ever suspect a missed event, the Webhooks API lets you replay any delivery within 30 days.

INC-2026-0174 — Decisioning timeouts from bureau upstream

Minor   Decisioning   ·   3 Jun 2026 · duration 1 h 06 min

• Impact — A commercial-credit bureau returned intermittent 504s. A subset of applications were returned as referred with reason data_incomplete instead of an instant decision; none were wrongly declined.
• Root cause — Upstream provider outage. Our in-request timeout (4s) tripped the graceful-degradation path as designed, holding applications for manual review rather than failing them.
• Follow-up — Added a secondary bureau as automatic fallback; referred applications from the window were re-scored and cleared within the hour.

INC-2026-0166 — OAuth token endpoint 5xx spike

Major   OAuth / OIDC   ·   21 May 2026 · duration 22 min

• Impact — The token endpoint returned 503s for ~22 minutes; new access tokens could not be minted, so fresh API sessions failed. Existing valid tokens continued to work, so most live integrations were unaffected.
• Root cause — Connection-pool exhaustion against the signing-key store after a traffic surge. Pool sizing was raised and a circuit breaker added in front of the key store.
• Follow-up — Token-mint latency now has its own SLO and page; we recommend clients cache access tokens until expiry rather than minting per request (see OAuth 2.0).

Stay informed

We post to this page first, then update by email and webhook. For money-movement-critical integrations we strongly recommend subscribing the status webhook so incidents surface inside your own tooling.

summary_json

GET https://status.credicorp.co.uk/api/v1/summary — public, no auth, cache for 30s.

webhook_events

status.incident.opened, status.incident.updated, status.incident.resolved.

email

Subscribe at status.credicorp.co.uk for per-incident email.

history

Incidents and uptime are retained and queryable for 90 days.