Developer Latency Sensitivity & SLO Benchmarking Survey

Welcome, and thank you for your interest in this survey on developer latency experiences. This survey takes approximately 11 minutes. Your participation is entirely voluntary, and you may stop at any time. There are no right or wrong answers — we are interested in your honest opinions and real-world experiences from the last 30 days. All responses are confidential, will be anonymized, and reported only in aggregate for internal research purposes.

Have you written, reviewed, or deployed code in a professional capacity in the last 30 days?

• Yes
• No

Which of the following languages or platforms did you actively use in the last 30 days? (Select all that apply)

• JavaScript/Node.js
• TypeScript
• Python
• Java
• Go
• Rust
• .NET/C#
• Ruby
• Kotlin
• Swift
• C/C++
• Other (please specify)

Overall, how sensitive to latency are your primary workloads?

Over the last 30 days, what p95 latency have you typically observed for your primary endpoint?

• < 50 ms
• 50–100 ms
• 100–250 ms
• 250–500 ms
• 500 ms – 1 s
• 1–2 s
• 2–5 s
• > 5 s
• I don't monitor this metric

If your median latency meets its target, how acceptable are occasional latency spikes?

For a latency-sensitive workload, rank the following priorities from most to least important.

1. Median latency (p50)
2. Tail latency (p95/p99)
3. Availability/reliability
4. Cost efficiency
5. Throughput
6. Feature completeness
7. Developer productivity

We'd like to explore your latency trade-off decisions in a bit more depth. An AI moderator will ask you a couple of follow-up questions.

Which of the following best describes your current role?

• Backend engineer
• Frontend/web engineer
• Full-stack engineer
• Mobile engineer
• ML/AI engineer
• SRE/DevOps
• Data engineer
• Engineering manager
• Other (please specify)

Thank you for completing this survey! Your responses will be used in aggregate to help set better latency benchmarks and improve developer tooling experiences. If you have questions, please contact the research team.

Which of the following use cases are most relevant to your current work? (Select all that apply)

• User-facing web API
• Interactive UI actions
• Search/query
• Payments/auth/checkout
• Online ML inference
• Batch ML/offline scoring
• Streaming/real-time feeds
• Data pipelines/ETL
• Background jobs
• Build/test/dev tooling
• Other (please specify)

For user-facing requests, what do you consider an acceptable median (p50) latency?

• < 20 ms
• 20–50 ms
• 50–100 ms
• 100–200 ms
• 200–500 ms
• 500 ms – 1 s
• > 1 s

What is your typical default timeout setting for external API or service calls?

• < 500 ms
• 500 ms – 1 s
• 1–3 s
• 3–5 s
• 5–10 s
• 10–30 s
• > 30 s
• No explicit timeout set

When latency threatens your SLA or SLO, rank your top strategies in order of priority (drag to reorder).

1. Degrade non-critical features
2. Cache more aggressively
3. Precompute or batch work
4. Parallelize or partition requests
5. Return partial results
6. Scale up/out resources
7. Fail fast with retry/backoff

In your experience, above what latency do interactive actions start to feel noticeably slow to users?

• 100 ms
• 200 ms
• 300 ms
• 500 ms
• 800 ms
• 1 s
• > 1 s

Based on your responses in this survey, please share any additional thoughts about acceptable latency, tail behavior, or how latency considerations shape your system designs.

How many years of professional software development experience do you have?

• < 1
• 1–2
• 3–5
• 6–9
• 10–14
• 15+

For user-facing requests, what do you consider an acceptable 95th-percentile (p95) latency?

• < 50 ms
• 50–100 ms
• 100–250 ms
• 250–500 ms
• 500 ms – 1 s
• 1–2 s
• > 2 s

What is the maximum acceptable end-to-end latency you would set for interactive UI actions (e.g., button clicks, navigation)?

• < 100 ms
• 100–200 ms
• 200–500 ms
• 500 ms – 1 s
• 1–2 s
• > 2 s

Approximately how large is your organization?

• 1 (just me)
• 2–10
• 11–50
• 51–200
• 201–1,000
• 1,001–5,000
• > 5,000

How important is reducing tail latency (p95/p99) compared to reducing average latency for your workloads?

What is the maximum acceptable end-to-end latency you would set for synchronous API calls (e.g., REST/gRPC)?

• < 100 ms
• 100–250 ms
• 250–500 ms
• 500 ms – 1 s
• 1–3 s
• > 3 s

In which region are you primarily located?

• North America
• Latin America
• Europe
• Middle East
• Africa
• Asia
• Oceania
• Prefer not to say

What is the maximum acceptable end-to-end latency you would set for batch or background jobs?

• < 1 s
• 1–5 s
• 5–30 s
• 30 s – 2 min
• 2–10 min
• > 10 min