PEHAMLIMITED
All updates
Update

Chase Outage Update: Software Bug Slows US Banks While ChatGPT and Microsoft 365 Also Stumble

Chase blamed a software bug for the November 16 bank outage. Here is what happened, why major outages keep happening, and how to prepare.

17 Nov 2026Team Peham#DevOps#Security
chase bank outage update by Peham

Update, November 17, 2026: this article covers outage events reported up to that date and will be refreshed as new details emerge.

On Monday, November 16, customers of JPMorgan Chase, Bank of America, Wells Fargo and US Bank flooded Downdetector with reports of failed logins, stuck transfers and balances that would not load. Chase said the cause was a software bug in an on-premise hardware system and that it was not cyber-related.

It was not the only incident this month. ChatGPT logged several error spikes in early November, and Microsoft Outlook had its own disruption for some users in Europe, the Middle East and Africa. Each had a different trigger, but the same pattern sits underneath: one small failure inside a large, connected system.

This guide explains what happened, why major outages keep happening, the hurdles they create for people and businesses, and what you can do before the next one.

What happened: the major outages of November 2026

Chase and other US banks (November 16)

Reports built through Monday afternoon and spread across several banks. Users described trouble with online access, transfers and ATMs. Some Chase customers saw a message saying some services were not available and to try again later.

JPMorgan Chase said customers still had access to their accounts and funds, and that the bug was slowing some transactions. A spokesman added that the issue was not systemic and that other banks were affected only where they were on the other side of a transaction with Chase. That detail matters, because it shows how one bank's outage can slow another bank without that second bank doing anything wrong. Zelle also saw a rise in outage reports, and US Bank said it had no problems of its own even though Downdetector showed user reports for it.

ChatGPT and OpenAI (November 4 to 11)

  • November 4: elevated error rates for paid ChatGPT plans, lasting about two hours.
  • November 5 to 6: errors affecting ChatGPT Work and Dots, rated minor, with a recorded duration of about 13 hours.
  • November 11: a short outage of around 10 minutes affecting logins, sign-ups and several ChatGPT features.

None of these was a full shutdown. Partial failures are common, and they still break workflows for anyone who depends on the service.

Microsoft 365 and Outlook

Microsoft 365 has had a rough year. A January 22 incident disrupted Outlook and other services across North America for close to 10 hours. On August 31, Microsoft pointed to a core authentication configuration shared by several Microsoft 365 services. On November 9, some users in Europe, the Middle East and Africa could not open mailboxes in Outlook on the web for roughly four and a half hours.

Why major outages happen

Big outages rarely start with one dramatic event. They usually start with ordinary things going wrong at scale.

1. A change goes out and something breaks

In an Uptime Institute survey, configuration and change management failures caused almost half of network-related outages. Cloudflare's November 2025 incident is a clear example: a change caused a configuration file used by its Bot Management system to double in size unexpectedly. On October 29, 2025, a configuration change to Azure Front Door disrupted Office 365, Teams, Xbox Live and Alaska Airlines for about nine hours.

2. Hidden dependencies

On October 20, 2025, a defect in an AWS DynamoDB update caused a DNS failure that cascaded to 113 services and lasted more than 15 hours. Many affected companies had never planned for that one service failing. The Chase outage followed the same logic on a smaller scale: other banks slowed down because they were handling transactions with Chase.

3. Software bugs and legacy hardware

Chase named a software bug in an on-premise hardware system. Large banks often run core processing on infrastructure built over many years, with newer apps layered on top. When the foundation slips, every app above it feels it.

4. Human error and weak procedures

Uptime data cited by Opengear says 85% of outages caused by human error come from staff not following procedures or from flaws in the procedures themselves. InfoWorld's analysis of cloud outages calls stronger runbooks, training, failure drills and operational guardrails increasingly central to resilience.

5. Power, networks and the physical world

Power remains the leading cause of impactful outages in Uptime's tracking, and IT and networking issues made up roughly 23% of them in 2024. Physical risk is growing too. Cloudflare's Q2 2026 report notes that AWS's me-central-1 region in the UAE was damaged by conflict in the Middle East.

The hurdles people and businesses face during an outage

For individuals

  • No access to balances, transfers or bill payments, which can lead to late fees.
  • Alarming screens, such as $0 balances or "temporarily unavailable" messages, as seen in past Bank of America outages.
  • Little information. Banks often confirm a problem without explaining the cause or giving a timeline.
  • Scam risk, as people search social media for help and run into fake support accounts.

For businesses

  • Email, meetings and file access stop together when one identity or authentication layer fails.
  • Your product breaks when a supplier breaks. Unite.AI noted that an OpenAI outage becomes the outage of every team selling software built on those calls.
  • Real cost. In Uptime's survey, 54% of organizations reported a recent outage that cost more than $100,000.
  • Heavy support load and lost trust while customers wait for answers.

What to do when your bank or app goes down

  1. Confirm it is the bank. Check Downdetector and the bank's official status page or verified social account before resetting any passwords.
  2. Switch channels. The website may work when the app does not, or the reverse. Phone banking is another route.
  3. Use your cards and ATMs. Card networks run separately from your bank's own systems, so debit and credit cards often keep working.
  4. Keep key details outside the app. Save the bank's phone number, plus your routing and account numbers, somewhere you can reach without logging in.
  5. Hold a small cash reserve for essentials.
  6. Ask for fee refunds. If the outage caused an overdraft or late fee, contact your bank. Some banks refund automatically, as Wells Fargo did after a major 2019 outage.
  7. Watch for scams. Contact support only through official channels.

How businesses can prepare for the next outage

You cannot stop Chase, OpenAI or Microsoft from failing. You can decide how much damage it does to you. These are the practices worth putting in place.

  • Map your dependencies. List every external service your product needs: payments, identity, AI APIs, email, DNS and hosting. Mark the ones with no fallback.
  • Design for graceful failure. If an AI API times out, queue the request, serve a cached answer or show a clear message instead of a blank screen.
  • Roll out changes in stages. Canary releases, feature flags and fast rollbacks limit how far a bad change can travel.
  • Spread critical paths. Use multiple availability zones or regions for the parts that must stay up, and consider a second provider for the most critical ones.
  • Monitor from outside. Track uptime and error rates independently of your own cloud, and keep a public status page.
  • Rehearse. Write runbooks, then run failure drills so the team has done it before the real incident.

If you want help with any of this, our Cloud and DevOps team designs and runs resilient infrastructure. Our custom software development and AI agent development teams build products that keep working when a dependency does not.

Not every outage is an attack, but patching discipline and change control reduce both kinds of risk. For related reading, see how PHP2Shell attacks targeted US banks, our November 2026 security roundup and our step-by-step PHP2Shell patch guide.

FAQ: bank, ChatGPT and Microsoft outages

Was the Chase outage a cyberattack?

Chase said it was not. The bank attributed the problem to a software bug in an on-premise hardware system and said it was not cyber-related.

Is my money safe during a bank outage?

In this case, Chase said customers still had access to their accounts and funds. An outage usually affects access and display, so check your balance again once service returns and report any mismatch to the bank.

Why do outages hit several companies at once?

Companies share cloud providers, payment networks, identity services and DNS. When one shared layer fails, everything built on it feels it, as the AWS and Azure incidents in 2025 showed.

How long do major outages last?

It varies widely. The incidents covered here ran from about 10 minutes to more than 15 hours, and some partial failures linger long after the main fix.

How can a business reduce its outage risk?

Map your dependencies, build fallbacks for the ones that matter, release changes in stages and rehearse your recovery plan. A short review of your stack usually finds the single points of failure quickly.

Find your single points of failure before the next outage

Not sure where your product is exposed? Book a free 30 minute discovery call with Peham and we will walk through your stack, your dependencies and your recovery plan. You can also reach us through the contact page.

Book a free 30 minute call

Sources