hardScenarioDesignMidSeniorStaff
Traffic is a hundred times normal and some of it is real customers. What do you drop first?
A DDoS or 100x traffic spike needs load shedding that separates real users from abusive or low-value work. Absorb volume at the edge, serve cacheable requests, protect checkout paths and reject expensive new work quickly before queues amplify retries.
5 min readUpdated 2026-08-15
Answer
Short answer
During a 100x traffic spike, first absorb obvious volumetric traffic upstream, then preserve the highest-value user journeys. Shed new or expensive work before in-progress checkout work, keep queues short, return fast retry hints, and measure by funnel health rather than raw request count.
Was this answer helpful?
Likely follow-ups
- Your mobile app retries on 503 with no backoff. What does that do to the arithmetic, and what do you ship to fix it?
- Which requests would you keep serving if you had to cut capacity to a tenth for six hours, and who signs off on that list?
- A challenge page stops the flood and also stops your partner API integrations. How do you let them through without handing an attacker the same door?
- How would you know, an hour in, whether the traffic is an attack or an unannounced marketing campaign?
Related questions
- The trigger for the outage is gone and you have restarted the service, and it collapses again within a minute of taking traffic. Why will it not recover on its own?hardAlso on load-shedding and admission-control7 min
- You need to enforce 100 requests per minute per API key across a fleet of forty servers. Which limiting algorithm, and where does the counter live?hardAlso on load-shedding and admission-control8 min
- A dependency that normally answers in 80ms starts taking eight seconds. What in your service reacts, and in what order?hardAlso on backpressure7 min
- A job queue is backed up four hours and half the jobs are now pointless. What do you drain, and what do you drop?hardAlso on backpressure6 min