Fastly Certification Interview Questions and Answers (2027)
Fastly certification interview questions: what Fastly publishes about its badge and partner certification, plus VCL, purging, shielding and Compute Q&As.
Fastly's published credentials are small and specific: a free Edge Computing 101 badge for building a first Fastly Compute service, and a self-guided partner certification track for implementing Fastly security products. Fastly does not publish exam codes, question counts, passing scores or prices for these on fastly.com. This page explains what is published and gives 34 knowledge-check questions that match the skills each credential and a Fastly interview test.
It is the certification member of our Fastly set. For a study schedule, use the Fastly interview preparation guide. For entry-level product questions, read Fastly engineer interview questions for freshers, and for incident practice, scenario-based Fastly interview questions. Facts were checked on fastly.com and Fastly's own community and investor sites in November 2026.
Key takeaways
- The Edge Computing 101 badge was announced on Fastly's community forum in December 2024. The post calls it completely free and aimed at beginners deploying a site to Fastly Compute.
- Fastly's channel partner page describes a technical certification program that trains partners to implement and manage Fastly security solutions as a partner-delivered service. Fastly announced its Certified Services Partner Program in November 2025.
- No exam code, price, duration or passing score is published on fastly.com for either credential. Treat any third-party "Fastly certification dump" as unofficial.
- The skills behind both are the same ones interviews test: VCL request flow, caching and purging, shielding, Fastly Compute with WebAssembly, and security products such as Next-Gen WAF.
What Fastly publishes about its credentials
| Credential | Who it is for | What Fastly says | Exam details published |
|---|---|---|---|
| Edge Computing 101 badge | Beginners and developers | Earned by deploying a website to Fastly Compute through a guided learning path; free | None |
| Partner technical certification (Certified Services Partner Program) | Technical staff at Fastly partner organisations | Self-guided training to implement and manage Fastly security solutions; announced November 2025 | None on fastly.com; contact Fastly's partner team |
If an employer asks whether you are "Fastly certified", the honest answer is to name the badge or partner credential you hold, if any, and then show the skills. Most Fastly interviews care more about whether you can read VCL and explain a cache miss than about a badge.
Fresher questions (Edge Computing 101 level)
1. What is Fastly?
Fastly is an edge cloud platform. It caches and delivers content from points of presence (POPs) close to users, runs customer code at the edge with Fastly Compute, and offers security products such as DDoS protection and Next-Gen WAF.
2. What is a Fastly service?
A service is the configuration unit: domains, backends (origins), caching rules and logic. Services are either VCL services, configured with Fastly's Varnish-based configuration language, or Compute services, which run your compiled WebAssembly program.
3. What is Fastly Compute?
Fastly's edge compute platform. Your code is compiled to WebAssembly and run with Wasmtime on Fastly's network. Fastly lists four official SDKs: Rust, JavaScript, Go and C++.
4. Why WebAssembly for edge code?
Fastly's docs give security and portability as the reasons: each request runs in an isolated sandbox, and the same compiled package runs anywhere on the network regardless of the language you wrote it in.
5. Which CLI commands create and deploy a Compute project?
fastly compute init creates a project, fastly compute build compiles it, fastly compute serve builds and runs it locally, and fastly compute publish builds and deploys in one step. fastly compute deploy uploads an existing package.
fastly compute init
fastly compute serve
fastly compute publish
6. What does fastly compute serve use to run code locally?
A local test server called Viceroy, which fastly compute install-tools installs along with the other tools the compute subcommands need.
7. What is a backend?
The origin server Fastly fetches content from on a cache miss or pass, such as a cloud storage bucket or an application load balancer.
8. What is a cache hit ratio?
The share of requests served from cache instead of the origin. A higher ratio means less origin load and lower latency for users.
9. What is a POP?
A point of presence: a Fastly data centre location with caching servers. Users connect to a nearby POP, which serves cached content or fetches from the origin.
10. How does Fastly decide how long to cache an object?
From the origin's caching headers (Surrogate-Control, Cache-Control s-maxage or max-age, Expires) or rules you set in VCL. In the VCL boilerplate, when none of those headers is present, the default TTL is 3600 seconds.
Questions for 2 to 5 years of experience (VCL and caching)
11. List the VCL subroutines in execution order.
Fastly's boilerplate presents them in this order: vcl_recv, vcl_hash, vcl_hit, vcl_miss, vcl_pass, vcl_fetch, vcl_error, vcl_deliver, vcl_log. Not every request runs every one; a hit skips miss and fetch.
12. Trace a cache miss through VCL.
vcl_recv returns lookup, vcl_hash builds the cache key, the lookup misses so vcl_miss runs and returns fetch, the origin responds and vcl_fetch decides whether to cache, then vcl_deliver sends the response and vcl_log runs logging.
13. Which requests does the boilerplate pass by default?
vcl_recv returns pass for methods other than GET, HEAD and FASTLYPURGE. In vcl_fetch, responses with Set-Cookie, or with Cache-Control containing private or no-store, are passed rather than cached.
14. Which subroutine is the only one that can return fetch?
vcl_miss. Fastly's reference notes that a custom subroutine returning fetch can therefore only be called from vcl_miss.
15. What rules apply to custom subroutines?
They use the sub keyword and cannot use the reserved vcl_ prefix, each name can be declared only once, and subroutines cannot call themselves directly or indirectly. Built-in subroutines with the same name are concatenated in the order the compiler reads them.
16. What purge types does Fastly support?
URL purge (one object), surrogate key purge (every object tagged with a key) and purge all (the whole service). URL and surrogate key purges can be soft or hard; purge all is always hard.
17. What is a soft purge and when do you use it?
A soft purge marks objects stale instead of removing them, so the next request can be served the stale copy while Fastly fetches a fresh one. Use it to protect the origin from a stampede after a content update.
18. How do surrogate keys help a news or e-commerce site?
The origin tags responses with keys in the Surrogate-Key header, such as a product ID and a category. When a product changes, one purge on its key clears every page that shows it. Each key counts as one purge however many objects it tags.
19. What are the purge limits?
Fastly's docs say URL and surrogate key purges do not count toward global API rate limits but are capped at an average of 100,000 per hour per customer, while purge all counts toward API rate limits.
20. Show the CLI command to soft purge a surrogate key.
fastly service purge --key product-42 --soft --service-id <SERVICE_ID>
Other flags are --url, --all and --file for a newline-delimited list of keys.
21. What is shielding?
Shielding routes requests from all edge POPs through one designated shield POP before the origin, so misses at the edge can be answered from the shield's cache. It reduces origin requests and improves the overall cache hit ratio.
22. What should you watch out for with shielding?
Your edge code usually runs twice, at the edge POP and at the shield, so place client-facing and origin-facing logic carefully. Reported hit ratio can look lower because a request that misses at both counts as two misses. Shielding cannot be applied to backends defined in VCL and is not available in the Compute C++ SDK.
23. How do you choose a shield POP?
Pick one close to the origin, taking into account interconnects with your cloud provider and the POP's cache capacity, as Fastly's shielding guide advises.
Senior questions (security and architecture)
24. What does the partner security certification focus on?
Fastly's channel partner page describes training to implement and manage Fastly's advanced security solutions as a partner-delivered service. Coverage reported at launch included DDoS protection, Next-Gen WAF and client-side protection; confirm the current scope with Fastly's partner team before you study.
25. What is Fastly Next-Gen WAF?
Fastly's web application firewall. Its documentation is organised around workspaces (also called sites), signals, rules and thresholds, and it can be deployed in different ways, including at the edge and with an agent in your own environment.
26. How would you roll out a WAF without blocking real users?
Start in a logging or monitoring mode, review which requests would be blocked, tune rules and exclusions for false positives, then switch to blocking for the most reliable rules first. Keep alerting on block rate after go-live.
27. How does a CDN help with DDoS?
A large distributed network absorbs volumetric traffic across many POPs, caching serves repeated requests without touching the origin, and rate limiting or WAF rules block abusive request patterns. The origin should accept traffic only from the CDN.
28. How do you protect the origin from being bypassed?
Restrict origin access to Fastly, for example with firewall allow-lists of Fastly's published IP ranges, a shared secret header checked by the origin, or mutual TLS between Fastly and the origin.
29. VCL or Compute for a new project?
VCL suits caching logic, header changes, redirects and routing, and has a large body of examples. Compute suits logic that needs a general-purpose language, libraries or more complex processing. Teams often keep caching in VCL and add Compute where custom code is needed.
30. How do you deploy Fastly configuration safely?
Keep configuration in Git, use the Fastly CLI or Terraform provider in CI, validate changes on a staging service, and activate new service versions with a quick rollback path to the previous version. Fastly services are versioned, which makes rollback a version activation.
Scenario questions
31. After a deploy, the cache hit ratio drops from 90 percent to 40 percent.
Check whether the origin started sending Set-Cookie or Cache-Control private on cacheable pages, which the boilerplate passes. Check for a new query string or header that changed the cache key in vcl_hash. Compare responses before and after with response headers such as X-Cache.
32. A product price change is not showing for some users.
Confirm the page carries the product's surrogate key, purge that key (soft purge if origin load is a concern), and check whether a browser or intermediate cache is holding the old copy because of a long max-age sent to clients.
33. Origin load spiked after a purge all.
Purge all is a hard purge, so every POP refetched at once. Prefer surrogate key purges for targeted changes, use soft purges, and enable shielding so edge POPs refetch from the shield instead of the origin.
34. A Compute service works locally with fastly compute serve but fails after publish.
Compare backend definitions and hostnames between fastly.toml local settings and the deployed service, check that secrets or config stores used in production exist, and read the service's logs through a logging endpoint. Run fastly compute validate on the package.
How to prepare
- Earn the Edge Computing 101 badge by following Fastly's learning path.
- Create a free Fastly account, set up a VCL service in front of a test origin, and practise URL, key and soft purges with the CLI.
- Write a small Compute app in JavaScript or Rust and deploy it with
fastly compute publish. - Read the VCL subroutine reference and the shielding guide until you can draw the request flow from memory.
Related: Fastly FAQs asked in DevOps interviews, Akamai CDN interview questions for comparison, and where CDN integrations help DevOps deployments. Edge work builds on networking and CI/CD skills covered in our DevOps training in Kharadi, Pune and live online.
Frequently asked questions
Does Fastly have an official certification exam?
Fastly publishes an Edge Computing 101 badge and a partner technical certification program for security implementation. It does not publish exam codes, question counts, passing scores or prices for them on fastly.com.
Is the Fastly Edge Computing 101 badge free?
Yes. Fastly's community announcement in December 2024 described it as completely free and easy for beginners to earn.
Who can take the Fastly partner certification?
It is designed for technical staff at Fastly partner organisations. Contact Fastly's partner team for eligibility and current requirements.
What topics should I study for a Fastly role?
VCL request flow, caching headers, purging with surrogate keys, shielding, Fastly Compute with WebAssembly, the Fastly CLI and security products such as Next-Gen WAF.
Which languages does Fastly Compute support?
Fastly lists official SDKs for Rust, JavaScript, Go and C++. Community SDKs exist for other languages.
What is the difference between soft and hard purge?
A soft purge marks content stale so it can still be served while refreshed. A hard purge removes it, so the next request must go to the origin.
Are online Fastly certification dumps reliable?
No. Fastly has not published an exam bank, so such material is unofficial. Build the skills on a real service instead.
Official references
What's Your Reaction?
Like
0
Dislike
0
Love
0
Funny
0
Angry
0
Sad
0
Wow
0