Angielski dla Cloud Architekta — słownictwo AWS, Azure i GCP
Architecture Review Board, zagraniczny CTO, pytanie o failover między regionami — a Ty szukasz słów. Ten artykuł daje Ci vocabularę do obrony architektury, zarządzania incydentami i rozmów o kosztach chmury po angielsku.
Twoja architektura jest świetna — ale nie umiesz jej obronić po angielsku
Architecture Review Board. Zagraniczny CTO zadaje pytanie: "How do you plan to handle failover between regions, and what's the expected RTO?" Twoja architektura jest przemyślana, trade-offy są uzasadnione — ale szukasz słów i tracisz wiarygodność.
Jeśli pracujesz jako Cloud Architect, Solutions Architect, DevOps Engineer, Platform Engineer lub Cloud Engineer, ten artykuł daje Ci vocabularę, której potrzebujesz na Architecture Review, incident calu i rozmowie o kosztach z management.
Infrastruktura — 10 terminów
| EN Term | PL | Przykład zdania |
|---|---|---|
| virtual machine (VM) | maszyna wirtualna | "We're migrating from bare-metal servers to virtual machines to improve resource utilisation." |
| container | kontener | "Each microservice runs in its own container with isolated dependencies." |
| Kubernetes (K8s) | Kubernetes | "We use Kubernetes to orchestrate container workloads across three availability zones." |
| microservices | mikroserwisy | "We decomposed the monolith into twelve microservices to enable independent deployments." |
| serverless | bezserwerowy | "The image processing pipeline is fully serverless — we only pay per invocation." |
| load balancer | równoważnik obciążenia | "The load balancer distributes traffic evenly across all healthy instances." |
| auto-scaling | automatyczne skalowanie | "Auto-scaling kicks in when CPU utilisation exceeds 70% for more than two minutes." |
| availability zone (AZ) | strefa dostępności | "We deploy across three availability zones to ensure no single point of failure." |
| region | region | "The primary region is eu-west-1; the DR site is eu-central-1." |
| edge computing | przetwarzanie brzegowe | "We moved latency-sensitive processing to edge computing nodes closer to end users." |
Storage i sieć — 8 terminów
| EN Term | PL | Przykład zdania |
|---|---|---|
| object storage | magazyn obiektowy | "Static assets and backups are stored in object storage — cheaper than block storage at this scale." |
| block storage | magazyn blokowy | "The database requires block storage with provisioned IOPS for consistent low latency." |
| CDN | sieć dostarczania treści | "After enabling the CDN, average page load time dropped from 2.8s to 0.4s for European users." |
| VPC (Virtual Private Cloud) | wirtualna sieć prywatna | "All production workloads run inside a VPC with no direct public internet access." |
| subnet | podsieć | "Public-facing services sit in the public subnet; databases are in the private subnet." |
| firewall | zapora sieciowa | "The firewall rules allow inbound HTTPS traffic on port 443 only from whitelisted IPs." |
| bandwidth | przepustowość sieciowa | "We're hitting bandwidth limits on the inter-region peering connection during business hours." |
| DNS | system nazw domenowych | "DNS failover redirects traffic to the secondary region within 60 seconds of a health check failure." |
Serwisy, architektura i operacje — 10 terminów
| EN Term | PL | Przykład zdania |
|---|---|---|
| managed service | serwis zarządzany | "Using a managed service for the database removes the operational overhead of patching and backups." |
| SLA | umowa o poziomie usług | "Our SLA guarantees 99.9% uptime, which translates to less than 9 hours of downtime per year." |
| uptime | dostępność | "The platform has maintained 99.97% uptime over the past 12 months." |
| disaster recovery (DR) | odtwarzanie po awarii | "Our disaster recovery plan targets an RTO of 4 hours and an RPO of 15 minutes." |
| failover | przełączenie awaryjne | "Failover to the secondary region is automated and takes approximately 90 seconds end-to-end." |
| snapshot | migawka | "We take snapshots before every production deployment for rapid rollback." |
| infrastructure as code (IaC) | infrastruktura jako kod | "All infrastructure is defined as code in Terraform — no manual changes allowed in production." |
| CI/CD pipeline | potok CI/CD | "The CI/CD pipeline runs tests, builds the container image and deploys to staging automatically." |
| monitoring | monitorowanie | "We use a centralised monitoring stack — metrics, logs and traces in one place." |
| cost optimization | optymalizacja kosztów | "The FinOps review identified $12k/month in cost optimization opportunities." |
Scenariusze komunikacji
a) Architecture Review Board — 8 zwrotów
- "We chose a multi-region active-active setup to achieve 99.99% availability."
- "The trade-off here is cost versus resilience — active-active is 40% more expensive but eliminates the failover window."
- "All infrastructure is defined as code in Terraform, so provisioning is repeatable and auditable."
- "The single point of failure was the NAT gateway — we've now deployed one per availability zone."
- "We selected a managed database service to reduce operational overhead and inherit the provider's SLA."
- "The RTO is 4 hours and the RPO is 15 minutes, which aligns with the business continuity requirements."
- "Auto-scaling is configured to handle three times our current peak load without manual intervention."
- "Security groups enforce least-privilege network access — no resource has broader permissions than required."
b) Incident response — 6 zwrotów
- "We're seeing elevated error rates in us-east-1 — initiating failover to eu-west-1."
- "The root cause is a misconfigured security group blocking traffic on port 5432 to the database."
- "All services have been restored. The incident lasted 23 minutes — post-mortem will follow within 48 hours."
- "We're hitting connection pool limits on the primary database. Scaling up the instance type as an immediate mitigation."
- "The auto-scaling group is healthy but slow to react — we're adjusting the scale-out cooldown period."
- "I'm rolling back to the previous snapshot — estimated restore time is 12 minutes."
c) Rozmowa o kosztach z management — 6 zwrotów
- "By migrating baseline workloads to reserved instances we can reduce our monthly cloud spend by approximately 35%."
- "Our current monthly cloud bill is $48k — 60% of that is compute, 25% is data transfer."
- "Spot instances are appropriate for the batch jobs because they're fault-tolerant and not time-sensitive."
- "The FinOps review flagged unused resources costing $3,200 per month — I've scheduled them for decommission."
- "I'd recommend setting up budget alerts at 80% and 100% of the monthly threshold."
- "The cost increase is proportional to the traffic growth — unit economics are stable."
Krótki dialog — Cloud Architect przedstawia projekt CTO
CTO: "Walk me through the failover strategy. What happens if us-east-1 goes down entirely?"
Cloud Architect: "We run an active-passive setup. The secondary region in eu-west-1 has a warm standby that can take traffic within 90 seconds. RTO is 4 hours for a full regional failure, RPO is 15 minutes — both are within the business continuity requirements."
CTO: "What's the cost of maintaining that standby environment?"
Cloud Architect: "Roughly 30% of the primary region cost. We could reduce that with a cold standby, but we'd extend RTO to 6 hours. Given our SLA commitments, I'd recommend keeping the warm standby."
CTO: "Agreed. Document the trade-off in the ADR and get sign-off from the business continuity team."
Terminy specyficzne dla dostawców chmury — 8
| EN Term | Dostawca | Opis |
|---|---|---|
| EC2 | AWS | Wirtualne maszyny w chmurze AWS |
| S3 | AWS | Object storage — standard branżowy |
| Lambda | AWS | Serverless compute — funkcje na żądanie |
| RDS | AWS | Zarządzane bazy relacyjne |
| EKS | AWS | Zarządzany Kubernetes w AWS |
| Azure DevOps | Azure | CI/CD i zarządzanie projektami Microsoft |
| AKS | Azure | Zarządzany Kubernetes w Azure |
| Blob Storage | Azure | Object storage odpowiednik S3 |
Najczęstsze błędy Polaków
1. "server" zamiast "instance". W chmurze nie ma fizycznych serwerów — są instancje. ❌ "I'll restart the server." → ✅ "I'll restart the instance / pod / service."
2. "open the port" → zbyt skrótowe. ❌ "Can you open port 443?" → ✅ "Can you allow inbound traffic on port 443 from the load balancer's security group?"
3. "the cloud is down" → zbyt ogólne. ❌ "The cloud is down again." → ✅ "We're experiencing an outage in the us-east-1 region affecting EC2 and RDS services."
4. Wymowa i użycie SLA. SLA to akronim czytany jako trzy litery: ess-ell-ay. ❌ "We have a 99.9 SLA." → ✅ "Our SLA guarantees 99.9% uptime, which translates to less than 9 hours of downtime per year."
5. "I will make a backup". ❌ "I will make a backup." → ✅ "I'll take / run / create a backup before the deployment."
Quick Reference Table
| EN Term | PL Tłumaczenie | Kontekst |
|---|---|---|
| virtual machine | maszyna wirtualna | infrastruktura |
| container | kontener | infrastruktura |
| Kubernetes | Kubernetes | infrastruktura |
| microservices | mikroserwisy | architektura |
| serverless | bezserwerowy | architektura |
| load balancer | równoważnik obciążenia | infrastruktura |
| auto-scaling | automatyczne skalowanie | infrastruktura |
| availability zone | strefa dostępności | infrastruktura |
| region | region | infrastruktura |
| edge computing | przetwarzanie brzegowe | architektura |
| object storage | magazyn obiektowy | storage |
| block storage | magazyn blokowy | storage |
| CDN | sieć dostarczania treści | sieć |
| VPC | wirtualna sieć prywatna | sieć |
| subnet | podsieć | sieć |
| firewall | zapora sieciowa | bezpieczeństwo |
| bandwidth | przepustowość sieciowa | sieć |
| DNS | system nazw domenowych | sieć |
| managed service | serwis zarządzany | operacje |
| SLA | umowa o poziomie usług | governance |
| uptime | dostępność | governance |
| disaster recovery | odtwarzanie po awarii | operacje |
| failover | przełączenie awaryjne | operacje |
| snapshot | migawka | operacje |
| infrastructure as code | infrastruktura jako kod | DevOps |
| CI/CD pipeline | potok CI/CD | DevOps |
| monitoring | monitorowanie | operacje |
| cost optimization | optymalizacja kosztów | FinOps |
| reserved instance | instancja zarezerwowana | FinOps |
| spot instance | instancja spot | FinOps |
Podsumowanie
Terminologia chmurowa to dziś lingua franca każdego architekta. Niezależnie od tego czy bronisz projektu na Architecture Review Board, zarządzasz incydentem o 2 w nocy, czy przekonujesz CFO do przejścia na reserved instances — te zwroty robią różnicę.
Więcej kontekstu IT znajdziesz w artykule o słownictwie IT po angielsku. Jeśli pracujesz w środowisku programistycznym, przydatny będzie też Business English dla programistów.
Gotowe fiszki z terminologią Cloud znajdziesz w ścieżce Cloud Architect w sekcji IT & Programowanie.