Search the practice set
275 questions, 991 terms and 600 topics in 30 areas.
60 results for “Platform Team Topologies”
Enabling Team
A team whose output is other teams' increased capability, working alongside them temporarily and deliberately removing itself.
Platform Team Product Model
Running an internal platform team with product management disciplines — users, roadmap, support, metrics — rather than as an infrastructure function.
Team Topologies
A model of four team types and three interaction modes used to design an organisation deliberately for a target architecture and flow of change.
Team Interaction Mode
The defined way two teams work together — collaborating, consuming a service, or facilitating — chosen deliberately and expected to change over time.
Cognitive Load
The total amount a team must hold in its head to work effectively, and a real constraint on how many services or domains one team can own.
Showback
Reporting each team's share of platform and infrastructure cost without actually charging it, to change behaviour without creating a market.
Booking.com's Experimentation Platform
Booking.com runs over a thousand concurrent experiments and treats the ability to test any change safely as a platform capability rather than a product feature.
A platform team proposes adopting a service mesh for 40 services. Make the case for and against, then decide.
What it genuinely provides mTLS everywhere, with automated certificate rotation. This is usually the deciding factor. Doing mTLS by hand means short lived certi
Application teams say the platform is unreliable. The platform team's dashboard shows 99.95% on every component. How do you resolve this?
Both are right, and that is the finding The platform measured its components. Consumers experience journeys. The gap between those two views is where trust erod
You are forming a platform team of four to serve twelve product teams. What do you build first?
Not a platform. One golden path. Four people cannot build an internal developer platform for twelve teams. They can make one journey excellent, and expand from
Your API gateway configuration has grown to 8,000 lines with request transformations and cross-service orchestration. Changes require a platform team ticket and take two weeks. How do you fix this?
Name the failure The gateway has become a distributed monolith — shared, centrally owned, untestable in isolation, and on the critical path of every team's deli
Your company needs a feature flag and experimentation platform. A vendor charges a substantial annual fee; a team estimates six weeks to build one. Decide.
Apply the differentiation test first Is this a source of competitive advantage? Feature flagging is context, not core: customers do not choose you for it, and b
Your platform team is now a ticket queue — every environment, every permission, every new service goes through them. How do you get out of it?
Categorise the queue before automating anything Take a month of tickets and group them. The distribution is always uneven, and two or three categories are usual
Your platform team of five spends nearly all its capacity upgrading control planes across 38 Kubernetes clusters. How did this happen and how do you fix it?
How it happened There is no split axis, only history. Clusters were created per project, per experiment, per team, per acquisition — each for a reason that was
A CDO proposes moving to a data mesh because the central data team is a bottleneck with a nine-month backlog. How do you assess the proposal?
Agree with the diagnosis, examine the prescription The bottleneck is real and it is structural rather than a matter of capacity. A central team receives data fr
A team is choosing between Iceberg, Delta and Hudi for a new lakehouse. How would you approach the decision?
The formats are more alike than the debate suggests All three provide the same core: atomic commits over object storage, snapshots and time travel, schema evolu
A team proposes moving an API to edge functions to reduce latency for international users. What do you check before agreeing?
Where is the data? This is the question that settles most edge proposals. Compute at the edge that calls back to a single region database has moved the compute
A warehouse table now feeds the CRM through reverse ETL. It breaks, and the analytics team and the CRM team each say it is the other's problem. How do you resolve it?
The dispute is a design gap, not a personality problem A warehouse table built for analysis — nightly latency, best effort availability, schema changes made by
A business sponsor asks for a real-time data platform because "the competition has one". Reporting is currently a nightly batch that lands at 06:00 and nobody has complained. How do you handle this?
Do not answer the technology question "Real time platform" is a solution, and it has arrived without a problem attached. Answering it directly leads either to a
A design review presents a new event-driven platform. What cost questions do you ask before approving it?
What the interviewer is testing Whether cost is part of your architecture review or an afterthought handled by finance later, and whether you know the specific
A downstream team needs to react to order changes. The order service can publish events, or they can consume CDC from its database. Which, and why?
The recommendation: published events, with CDC as the mechanism if needed The distinction that matters is what the consumer becomes coupled to . CDC consumed di
A front-end team wants to replace their global state library because "state management is unmanageable". How do you evaluate the request?
Ask what is actually in the store The complaint almost always resolves to one distinction not having been made: server state versus UI state . Server state is d
A global SaaS product has a 200 ms p95 API budget. Users in Sydney see 900 ms against a single US-East deployment. The application team says the service responds in 40 ms. Who is right and what do you do?
Both are right, and that is the point The service does respond in 40 ms. The other 860 ms is network — and no amount of application profiling will show it, whic
A marketing site is a client-rendered single-page application. Organic traffic is poor and the team blames the search engine. What is your assessment?
The rendering choice is the problem A marketing page is identical for every visitor and highly cacheable. Rendering it in the browser pays for personalisation t
A new platform must serve a public partner API, three internal front-ends with different data needs, and high-volume service-to-service traffic. Choose the API styles and defend the choice.
Resist "pick one" These are three different problems with three different consumers. Standardising on one style optimises for architectural tidiness at the expe
A nightly job occasionally runs twice, producing duplicate charges. The team proposes a distributed lock. What do you say?
The first response A lock will reduce the frequency and will not eliminate it , and if the team believes otherwise they will stop looking for the real fix. The
A platform of 40 services has logs only, and incidents take hours to diagnose. Design the observability strategy and its rollout order.
Why logs alone fail at this size Logs answer "what happened in this service". They cannot answer "where did this request spend its time across twelve services",
A regulated client requires that no traffic between their data centre and your SaaS platform traverses the public internet. Design the connectivity and justify the cost.
Two distinct requirements hiding in one sentence Traffic must not traverse the public internet — a routing requirement. The client must be able to demonstrate i
A regulator asks whether customer data is encrypted. The team says yes, disks are encrypted. Is that a sufficient answer?
What disk encryption actually protects against Someone obtaining the physical medium or a raw storage snapshot. In a cloud context that means a provider employe
A release must rename a heavily used database column and ships tonight. The team proposes doing the rename in the deployment. What is wrong, and what do you propose?
What is wrong A rename is not additive. During a rolling deployment both versions of the application run simultaneously: one expects the old name, one the new.
A single deploy took down your monitoring platform. What happened, and how do you prevent a recurrence?
What almost certainly happened A high cardinality label was added to a metric. Each unique combination of label values is a separate time series, and cost scale
A team is about to launch a public GraphQL API. What must be in place before it goes live?
Query cost control — the security requirement An arbitrary query language exposed publicly means a client can construct a query that consumes unbounded resource
A team is launching a new service and asks what its SLO should be. How do you help them decide, and why is "99.99%" usually the wrong first answer?
What the interviewer is testing Whether you treat reliability as a cost benefit decision with a budget, or as a virtue to maximise. Why not four nines It costs
A team is splitting a monolith into services. How do you decide where the boundaries go?
The criterion is what changes independently Parnas's information hiding is the sharpest formulation: decompose around the decisions most likely to change , not
A team of fifteen wants to start a new product as microservices. How do you advise?
Ask what problem microservices are solving here The pattern addresses independent deployment by independent teams, independent scaling of unlike workloads, and
A team proposes CQRS for a CRUD admin panel because reads are slow. Is that the right call?
Why not CQRS here "Reads are slow" is a symptom with many causes, and CQRS addresses only one of them: the read and write models having genuinely different opti
A team proposes CQRS with event sourcing for a new service. How do you evaluate the proposal?
Separate the two, because they are independent decisions CQRS is separating the write model from the read model. Event sourcing is storing state as a sequence o
A team proposes circuit breakers on every downstream call. What would you add, and why is the breaker not the most important control?
Why the bulkhead matters more A circuit breaker acts after it has detected a pattern of failure. During the seconds before it trips — and while its threshold is
A team proposes exposing their service's database change stream via CDC so other teams can consume it, avoiding the work of building an event API. What is your assessment?
Name what is actually being proposed The proposal is to publish the service's internal schema as its integration contract. CDC does not emit domain events; it e
A team proposes moving personalisation to edge functions to cut latency. When does that work and when does it backfire?
The principle that decides it The edge is near the user and far from your data. Edge compute pays off for work that needs the request but not your state . The m
A team proposes rewriting a critical system from scratch, arguing the existing one is unmaintainable. How do you evaluate this?
Test the diagnosis before the prescription "Unmaintainable" usually means one of several different things, and they have different remedies: Nobody understands
A team proposes storing the customer's address on every order row "so order history is accurate". Is that denormalisation or a modelling error?
The distinction that matters It is neither, quite — it is a temporal modelling requirement being solved by accident. Denormalisation duplicates a fact for perfo
A team says "backups run nightly and are retained for 30 days". You have one hour to assess whether they could actually recover. What do you check?
The checks, in the order that finds problems fastest 1. When was a restore last performed, and how long did it take? The single most informative question. If th
A team shows you a design with eight services. Without knowing the domain, what questions tell you whether the boundaries are right?
What the interviewer is testing Whether you can evaluate a structure from its properties rather than needing to be a domain expert in every system you review. T
A team tells you backups run nightly and are retained for 30 days. What is missing from that answer?
What the interviewer is testing Whether you know that a backup is not a control until it has been restored — one of the most reliable ways to find an unrecovere
A team wants event sourcing for a new order service, citing audit requirements. What do you recommend?
The recommendation: probably an audit log, not event sourcing If the requirement is audit , event sourcing is a very expensive way to obtain it. An append only
A team wants to build a new internal API on serverless functions. It will serve steady traffic of about 200 requests per second during business hours. What do you advise?
What the interviewer is testing Whether you can apply the serverless trade off to a specific workload rather than treating it as a default good or a default bad
A team wants to move a workload from PostgreSQL to a document store because "the schema keeps changing". What do you ask?
The questions 1. What is actually changing — the shape, or the schema management process? "The schema keeps changing" usually means migrations are painful, not
A team wants to self-manage Kafka on Kubernetes to save money against the managed service. Evaluate.
The comparison they have probably made Instance cost versus managed service list price. That comparison usually favours self managing by a wide margin, and it o
A team's CI suite fails roughly one run in three for reasons unrelated to the change. Everyone reruns until green. How do you recover the situation?
Recognise what has actually been lost The suite is no longer a gate. Once the team's reflex on red is "rerun", that reflex is applied to genuine failures too, a
A team's pipeline takes 80 minutes from commit to a deployable artifact and engineers have stopped watching it. How do you approach fixing it?
Measure per stage before changing anything Get the duration of every stage over the last few hundred runs, with the distribution rather than the mean. The shape
A vendor claims their streaming platform provides exactly-once processing. How do you evaluate the claim?
Ask where the guarantee ends Nearly always at the platform's boundary. Within it, state and offsets commit together, so internal state reflects each input once.
An AI feature launched two months ago now costs more per month than the rest of the platform. What do you investigate?
Get cost per request, decomposed Token cost splits into input and output, and they price differently. Break the bill down by feature, by user, and by input vers
An e-commerce platform stores card numbers to support repeat purchases. How do you reduce PCI scope?
Stop the card number reaching your systems at all The strongest reduction is not storing cards more safely; it is never receiving them. A hosted field or an ifr
An end-to-end suite of 340 tests takes four hours and fails spuriously about half the time. The team wants to parallelise it. Is that the right move?
Parallelising treats the symptom It might halve the runtime. It will not touch the flakiness — in fact parallelisation often worsens it, by exposing shared stat
An internal AI assistant gives confidently wrong answers. The team wants to upgrade to a better model. What do you check first?
Establish whether the model ever saw the right content Log the retrieved chunks alongside each answer, then take the wrong answers and check: was the correct so
Platform Team Topologies
Stream-aligned, enabling, complicated-subsystem and platform teams, and their interactions.
Team Topologies
Stream-aligned, platform, enabling and complicated-subsystem teams.
Platform Funding
Central cost, showback, chargeback, and justifying a team that ships no customer feature.
Centralised vs Distributed
Shared platform leverage against team autonomy.