A major financial institution in North America experienced serious outages in its middleware application, which is essential for banking services. The Site Reliability Engineering (SRE) team analyzed thread dumps and found that a bug in the Oracle JDBC driver was causing too many active threads. Updating to a patched driver fixed the issue, stabilized operations, and helped prevent future outages.