SRE Capacity Planning Interview Problem: Handling Black Friday Traffic at an E‑Commerce Company
TL;DR
The correct answer is to demonstrate a data‑driven sizing process, a risk‑aware mitigation plan, and a clear communication framework. Interviewers reject vague “we’ll add servers” promises and reward concrete load‑model calculations. Your judgment must show that you can protect revenue during the Black Friday surge without over‑provisioning.
Who This Is For
You are a senior SRE or site reliability engineering lead who has already shipped a production service at scale. You are targeting senior‑level SRE roles at large e‑commerce firms, where the interview will focus on capacity planning for high‑traffic events. You likely earn a base salary between $190,000 and $220,000, with equity stakes of 0.04%–0.07% and a sign‑on bonus that can reach $30,000. You have at least three interview rounds: a phone screen, a system design deep dive, and a final on‑site debrief with the hiring manager and the capacity‑planning committee.
How should I estimate traffic volume for a Black Friday surge?
The judgment is: you must anchor your estimate in historical peak metrics, adjust for promotional lift, and validate with a short‑term load‑test. In a Q2 debrief, the hiring manager pushed back because I quoted a flat “2× traffic” without showing the derivation. I opened my notebook with the last year’s Black Friday page‑view curve, which peaked at 12 M requests per minute. I then added a 30 % uplift for a new “Lightning Deals” banner that had generated a 1.5× lift in the prior quarter. Finally, I ran a 48‑hour load‑test on a staging cluster, confirming that 15.6 M RPM sustained throughput was achievable with a 20 % headroom buffer.
First Insight – The “Historical‑Lift‑Validate” triad:
- Historical – pull the exact minute‑by‑minute traffic from the prior year.
- Lift – apply a factor based on known marketing campaigns, not generic “traffic spikes”.
- Validate – run a controlled load‑test to prove the model, even if the test runs for only two days.
The contrast is not “more servers will solve the problem,” but “right‑sized servers based on a calibrated model will solve the problem.”
Script for the interview:
> “I start with the exact 2022 Black Friday RPS, then I calculate the promotional uplift using Q3 campaign data, and finally I run a two‑day load‑test to confirm the sizing. That three‑step loop gave me a 15.6 M RPM target with a 20 % safety margin.”
What capacity‑planning framework survives the hiring manager’s debrief?
The judgment is: use the “Capacity‑Risk‑Mitigation” (CRM) framework, not a checklist of hardware specs. In a senior‑level interview, the capacity‑planning committee asked me to walk through my plan while the hiring manager listened for risk signals. I presented a single slide with three columns: Capacity (the 15.6 M RPM figure), Risk (single‑point‑of‑failure services, network saturation), and Mitigation (auto‑scaling policies, multi‑AZ redundancy). The hiring manager nodded when I said, “My risk is not the traffic volume, but the lack of cross‑region failover.”
Second Insight – Risk‑First Lens:
Most candidates treat capacity as a pure engineering problem. The counter‑intuitive truth is that risk, not capacity, drives the interviewer's decision. By foregrounding risk, you demonstrate that you think like a senior SRE who protects business outcomes.
Script for the interview:
> “My plan allocates 15.6 M RPM capacity, but I also identify two top‑risk buckets: the load‑balancer queue depth and the database connection pool. For each, I define a mitigation – a 30 second auto‑scale trigger and a read‑replica pool that expands to 1.5× the primary.”
How do I convince senior leadership that my plan is safe, not risky?
The judgment is: translate technical safety margins into business‑impact language, not just percentages. After the on‑site interview, the senior VP of Operations asked me to justify the 20 % headroom. I responded with a revenue‑impact story: “If we lose 0.5 % of transactions during peak, we forfeit $2.3 M in sales, which outweighs the $150 K cost of the extra capacity.” The hiring manager later told the committee that the candidate’s answer shifted the conversation from “Can we handle traffic?” to “Can we protect revenue?”
Third Insight – Business‑Impact Mapping:
Map every technical metric to a dollar impact. The interview is not about engineering elegance; it is about protecting the bottom line.
Script for the interview:
> “The 20 % buffer costs roughly $150 K in extra compute, but it reduces the probability of a revenue loss from $2.3 M to under $200 K, which is a net win for the business.”
Which metrics and alerts prove I’ve built a resilient system during the interview?
The judgment is: present a concise alert portfolio that covers latency, error rate, and resource saturation, not a generic “monitor everything” list. During the final debrief, the hiring manager asked for the three most critical alerts. I listed: (1) 99th‑percentile latency > 250 ms, (2) HTTP 5xx error rate > 0.2 %, and (3) CPU utilization > 75 % across all AZs. I explained why each alert aligns with a specific SLA breach scenario. The hiring manager noted that I avoided the “not enough alerts, but the right alerts” trap.
Fourth Insight – SLA‑Aligned Alert Set:
Select alerts that map one‑to‑one to Service Level Objectives. The interviewers reward a focused alert set that demonstrates understanding of failure modes.
Script for the interview:
> “My alerting strategy is threefold: latency spikes, error‑rate spikes, and CPU pressure. Each ties directly to a measurable SLA breach, so the on‑call team can prioritize remediation without noise.”
What post‑interview follow‑up demonstrates mastery of capacity planning?
The judgment is: send a concise debrief note that includes the refined capacity model, risk register, and next‑step recommendations, not a generic thank‑you email. Two days after the interview, I emailed the hiring manager a one‑page PDF titled “Black Friday Capacity Plan – Final Recommendations.” The document contained the updated 15.6 M RPM target, a revised risk matrix, and a rollout timeline: week 1 – staging validation, week 2 – canary in production, week 3 – full rollout. The hiring manager replied, “Your follow‑up shows you think beyond the interview room.”
Fifth Insight – Actionable Follow‑Up:
A senior SRE interview ends at the interview table, but the hiring committee continues to evaluate your post‑interview communication. A follow‑up that adds value is a decisive signal.
Script for the email:
> Subject: Black Friday Capacity Plan – Final Recommendations
> Body: “Hi [Hiring Manager], thank you for the discussion. Attached is the refined capacity model (15.6 M RPM), a risk register, and a three‑week rollout plan. I look forward to iterating on this with the team.”
Preparation Checklist
- Review the last two years of Black Friday traffic logs from the target company’s public metrics (if available) or a similar e‑commerce firm.
- Build a spreadsheet that calculates promotional lift using the “Historical‑Lift‑Validate” triad.
- Run a two‑day load‑test on a sandbox cluster to confirm the target throughput.
- Draft a one‑page “Capacity‑Risk‑Mitigation” slide that maps capacity, risk, and mitigation side by side.
- Translate each technical risk into a dollar impact statement, using the company’s average order value as a multiplier.
- Define a three‑alert portfolio aligned to SLA breach scenarios; test each alert in a staging environment.
- Work through a structured preparation system (the PM Interview Playbook covers capacity estimation with real debrief examples and includes a script library for post‑interview follow‑up).
Mistakes to Avoid
BAD: Saying “We’ll add more instances until the CPU is below 80 %.”
GOOD: Saying “I’ll add enough instances to sustain 15.6 M RPM with a 20 % headroom, verified by a load‑test, and I’ll monitor CPU at 75 % as an alert threshold.”
BAD: Claiming “The problem is traffic volume, so we need bigger servers.”
GOOD: Stating “The problem is capacity headroom; I model traffic, apply promotional lift, and design mitigation for the identified risk of load‑balancer queue overflow.”
BAD: Sending a generic thank‑you note after the interview.
GOOD: Sending a concise debrief that includes the refined capacity model, risk register, and a three‑week rollout plan, demonstrating continued ownership.
FAQ
What should I bring to the on‑site interview to prove I can size Black Friday traffic?
Bring a one‑page slide with the exact historical traffic curve, the promotional lift factor, and the validated target throughput. Include a brief risk matrix and the three SLA‑aligned alerts. The interviewers will judge you on the completeness of the data, not the number of slides.
How many interview rounds typically cover capacity planning for senior SRE roles?
Most senior SRE interviews have three rounds: a phone screen that checks fundamentals, a system‑design deep dive that includes the Black Friday scenario, and an on‑site debrief with the hiring manager and the capacity‑planning committee.
What salary can I expect if I ace this interview at a large e‑commerce firm?
A senior SRE who demonstrates the capacity‑risk‑mitigation approach can command a base salary between $190,000 and $220,000, an equity grant of 0.04%–0.07%, and a sign‑on bonus that may reach $30,000, depending on experience and market conditions.
The 0→1 PM Interview Playbook (2026 Edition) — view on Amazon →