Question
After a frontend deploy, a fraction of users keep getting the *old* JavaScript bundle and the app breaks for them with console errors about a missing API the new bundle expects, while other users are fine. The CDN dashboard shows high hit ratio (good) but you notice the deploy changed the cache-control on HTML from 'no-cache' to a 24h max-age 'to improve performance', while the JS bundles use content-hashed filenames. Some POPs and browsers are serving cached HTML that references old, now-deleted hashed bundles. How do you triage this partial breakage and fix the caching correctly?
Stop the bleeding first (mitigate), then form hypotheses from real signals. Separate root cause from symptom, communicate status as you go, and close with what prevents a repeat.