Update: Zelle. US Bank. JP Morgan. Bank of America. Wells Fargo Down too!
Shortly after 3:30 PM UTC, user reports of problems with Chase went from near zero to more than 2,000 on Downdetector. Chase processes trillions of dollars in daily transaction volume for tens of millions of digital users, so a spike like that is a serious availability failure. It also shows how hyperscale banking platforms break.
What Chase users reported during the outage
Users reported failed logins, account balances that would not load accurately, and transactions that would not complete. The problems appeared in both the mobile app and the web portal.
How Downdetector flagged the Chase outage
Downdetector is a crowdsourced monitor that handles millions of problem reports a month across tens of thousands of services. For each service it calculates a baseline: the average number of reports for that time of day, measured over a preceding period that is often six months to a year. It flags an incident only when report volume rises well above that baseline. It also limits the weight of repeat reports from the same user, so a spike reflects a geographically dispersed pattern and not one person refreshing the app.
For Chase, reports sat near zero through the morning and afternoon. Around 3:30 PM UTC they climbed almost vertically and went past 2,000. Against a baseline that low, the deviation is hard to dispute, and the signal may have appeared sooner than some internal monitoring tools would have registered the failure.

How Chase’s banking infrastructure is built to avoid outages
Chase.com supports over 67 million active users. Platforms at that scale are webs of microservices, distributed databases, and multi-cloud environments, not monoliths. The usual design is a multi-region active-active deployment, which keeps physical failures, bad weather, or localized cyber-attacks from taking everything down. Applications are grouped into zonal or regional deployments, and load balancing lets a single region carry the full workload if it has to.
That only works if applications are stateless. Conversational state moves into distributed caches, and services scale horizontally to absorb sudden jumps in requests per second. Global and regional DNS with health checks sits in front, along with CDNs that separate static assets from API calls and handle traffic at the edge so origin servers do not get swamped.
What likely caused the Chase outage
Downdetector shows when users noticed a failure and says nothing about the cause, so everything below is informed speculation.
Database, network, or API gateway failure
One candidate is a cascading failure in a core database cluster that handles account authentication. Another is a network partition that cuts communication between regional availability zones. A third is a widespread misconfiguration in the API gateway that governs traffic routing.
Third-party dependencies such as Plaid
Chase has previously had downtime affecting all channels and data recipients through API integrators like Plaid. An external service disruption can cascade into the core banking experience.
📬 Stay Ahead of Cyber Threats
Get the latest cybersecurity news, critical vulnerabilities, threat intelligence, tutorials, and exclusive giveaways delivered straight to your inbox. No spam. Unsubscribe anytime.
Users reported login errors alongside balance and transaction problems. That combination suggests the fault sat deep in the authentication and data retrieval pipeline, not in the front-end interface. Recovery from that kind of failure means site reliability engineers run root cause analysis, isolate the faulty component, and fail over to redundant systems. Depending on severity, that takes anywhere from minutes to hours.
Why the sudden spike points to a binary failure
JPMorgan Chase has described the Chase.com cloud migration as achieving resilience through multi-region isolation and automated “infrastructure repaving,” which removes security drift. These mechanisms are meant to detect anomalies and reroute traffic before users see degradation.
Capacity is one way they could be overrun. If the volume of simultaneous, legitimate requests exceeds planning parameters, perhaps after a volatile market event or a coordinated surge in user activity, graceful degradation gets tested to its limit. The shape of this spike argues against that. A sudden jump, as opposed to a gradual climb, points to a binary failure such as a DNS resolution timeout or an expired certificate. Resource exhaustion usually produces a more tapered rise in error rates.
What the outage data gives engineers
A sharp spike past 2,000 reports at 3:30 PM UTC gives engineering teams a precise timestamp. They can line it up against deployment logs, network traffic patterns, and database performance metrics to reconstruct the sequence of events that led to the degradation. Multi-region redundancy, stateless services, and edge computing lower the odds of an outage, but zero downtime remains a goal. No architecture guarantees it.
Chase outage FAQ
When did the Chase outage start?
According to Downdetector, reports rose from near zero to a steep climb at about 3:30 PM UTC.
How many people reported problems?
Downdetector reports passed 2,000 at the peak.
What were users unable to do?
Users reported login failures, inaccurate or missing account balances, and failed transactions in the Chase mobile app and web portal.
What caused the Chase outage?
The Downdetector data does not identify a cause. Plausible candidates are an authentication database failure, a network partition, an API gateway misconfiguration, a third-party dependency, a DNS timeout, or an expired certificate.
How does Downdetector decide something is an outage?
It compares current report volume against a baseline for that service and time of day, and flags an incident only when volume is significantly higher.









