Angielski dla DevOps Engineera — CI/CD, Kubernetes i pipeline po angielsku
Pipeline padł o 3 w nocy, on-call z Londynu pyta o status, a Ty szukasz słów na "failed deployment" i "CrashLoopBackOff". Ten artykuł daje Ci słownictwo do incident response, postmortem i codziennej komunikacji DevOps po angielsku.
Pipeline padł o 3 w nocy — i nie wiesz jak opisać failed deployment po angielsku
Środek nocy, alert w PagerDuty, on-call z Londynu pisze do Ciebie na Slack: "What happened? Is it a rollback scenario?" Ty wiesz dokładnie, co się stało — deployment poszedł na staging z błędną zmienną środowiskową, pod wchodzi w CrashLoopBackOff — ale brakuje Ci słów, żeby to przekazać szybko i precyzyjnie.
Taki scenariusz to codzienność DevOps Engineera w międzynarodowym środowisku. Z tego artykułu skorzystają: DevOps Engineer, Platform Engineer, SRE (Site Reliability Engineer), Infrastructure Engineer i Release Manager — wszyscy, którzy zarządzają pipeline'ami, infrastrukturą i deploymentami w środowiskach pracujących po angielsku.
CI/CD — 8 terminów
| EN Term | PL | Przykład w pracy |
|---|---|---|
| pipeline | potok CI/CD | "The pipeline failed at the integration test stage." |
| build | kompilacja / zbudowanie artefaktu | "The build took 12 minutes — can we parallelize the test steps?" |
| artifact | artefakt (wynik builda) | "The artifact from build #4217 was deployed to staging." |
| deployment | wdrożenie | "We use blue-green deployment to minimize downtime during releases." |
| rollback | wycofanie wdrożenia | "I'm initiating a rollback to version 2.3.1 — the error rate is still climbing." |
| staging environment | środowisko staging | "All changes must pass in staging before we promote to production." |
| production environment | środowisko produkcyjne | "Never test unverified changes directly in the production environment." |
| release | wydanie / release | "The release is scheduled for Friday 18:00 UTC after the freeze window." |
Infrastruktura i kontenery — 8 terminów
| EN Term | PL | Przykład w pracy |
|---|---|---|
| container | kontener | "Each microservice runs in its own container with defined resource limits." |
| image | obraz (kontenerowy) | "The Docker image is built from the base image and pushed to the registry." |
| pod | pod (jednostka Kubernetes) | "The pod is in CrashLoopBackOff — checking the logs now." |
| cluster | klaster | "We have three production clusters: EU-West, US-East and AP-Southeast." |
| node | węzeł klastra | "One node went NotReady — we drained it and rescheduled the workloads." |
| namespace | przestrzeń nazw | "All payment services run in the payments namespace with stricter network policies." |
| helm chart | helm chart (pakiet Kubernetes) | "Update the Helm chart values and push — the pipeline will handle the rest." |
| service mesh | service mesh | "We use Istio as the service mesh for mTLS between microservices." |
Monitoring i niezawodność — 8 terminów
| EN Term | PL | Przykład w pracy |
|---|---|---|
| uptime | dostępność systemu | "We achieved 99.95% uptime last quarter, just below our SLA target." |
| SLA (Service Level Agreement) | umowa o poziomie usług | "Our SLA with the client guarantees 99.9% availability per month." |
| SLO (Service Level Objective) | cel poziomu usług | "The SLO for checkout latency is p99 under 300ms." |
| SLI (Service Level Indicator) | wskaźnik poziomu usług | "Our key SLI is error rate — currently at 0.3%, well within the budget." |
| alert | alert / powiadomienie | "We're getting spurious alerts from the memory metric — let's tune the threshold." |
| on-call rotation | dyżur on-call | "I'm on-call this weekend — page me if error rate exceeds 1%." |
| postmortem | postmortem (analiza incydentu) | "The postmortem is scheduled for Tuesday — please add your timeline to the doc." |
| MTTR (Mean Time to Recovery) | średni czas naprawy | "Our MTTR improved from 47 minutes to 12 minutes after we automated the rollback." |
IaC i automatyzacja — 6 terminów
| EN Term | PL | Przykład w pracy |
|---|---|---|
| IaC (Infrastructure as Code) | infrastruktura jako kod | "All our infrastructure is managed as IaC — no manual changes in the console." |
| Terraform | Terraform | "The Terraform plan shows 3 resources will be modified and 1 destroyed." |
| Ansible | Ansible | "We use Ansible playbooks to configure the OS-level settings on new nodes." |
| idempotent | idempotentny | "Make sure the provisioning script is idempotent — it should be safe to run twice." |
| drift | dryfowanie konfiguracji | "There's configuration drift on the prod servers — someone made a manual change." |
| provisioning | inicjowanie / dostarczanie zasobów | "Provisioning a new environment takes under 10 minutes with our Terraform modules." |
Scenariusze komunikacji
a) Incident response — 8 zwrotów
- "We're seeing elevated error rates on the payment service — currently at 4.2%, up from a baseline of 0.1%."
- "I'm rolling back to the previous version now. ETA for rollback completion: approximately 3 minutes."
- "The pod is in CrashLoopBackOff — checking the application logs now."
- "Root cause is still under investigation. The working hypothesis is a misconfigured environment variable."
- "I've drained the affected node and rescheduled all workloads to healthy nodes in the cluster."
- "Error rate is back to baseline. The rollback was successful at 03:14 UTC."
- "I'll open a postmortem doc now — can everyone add their timeline of actions?"
- "This incident will be classified as SEV-2. Full postmortem to follow within 48 hours."
b) Omówienie deploymentu z zespołem dev — 6 zwrotów
- "The pipeline failed at the integration test stage. Can you check if this is related to your last commit?"
- "The build is green on the feature branch but failing on main — looks like a merge conflict introduced a regression."
- "We're planning a maintenance window this Friday between 22:00 and 23:00 UTC for the infra upgrade."
- "Please don't push to main during the release freeze — it runs from 16:00 today until we confirm production is stable."
- "The canary deployment is holding at 10% traffic. Error rates look good — I'll promote to 50% in 30 minutes."
- "Can you reduce the memory request in the Helm chart? The pod is getting OOMKilled on the smaller nodes."
c) Prezentacja postmortem — 6 zwrotów
- "Root cause was a misconfigured environment variable in the staging deployment that propagated to production."
- "The incident timeline shows a 23-minute gap between the first alert and the first human response — that's our biggest action item."
- "Contributing factors included an insufficient deployment checklist and no automated health check after rollout."
- "Action items: add environment variable validation to the pipeline, update the runbook, and review alert thresholds."
- "This incident cost us approximately 0.05% of our monthly SLA budget."
- "The key lesson learned is that we need to test configuration changes in an environment that mirrors production more closely."
Krótki dialog — DevOps Engineer na incident calu z team leadem
Team Lead: "Status update — what are we looking at?"
DevOps Engineer: "We have a SEV-2 incident on the checkout service. Error rate hit 6% at 03:08 UTC. I've identified the likely cause — the last deployment pushed an incorrect database connection string to the production config."
Team Lead: "Are you rolling back?"
DevOps Engineer: "Rollback is in progress now — should complete in about two minutes. I've also scaled down the affected pods to stop new requests hitting the broken version."
Team Lead: "Good. Keep me posted. What's the user impact?"
DevOps Engineer: "Roughly 15% of checkout attempts failed during the 6-minute window. Error rate is already dropping — we're back at 0.8% and falling."
Team Lead: "Confirmed stable, start the postmortem doc. I'll notify the stakeholders."
Postmortem — 6 kluczowych zwrotów ze struktury dokumentu
- Incident timeline: "23:47 UTC — anomalous error rate detected. 23:49 UTC — on-call engineer paged. 23:52 UTC — root cause identified. 23:55 UTC — rollback initiated."
- Root cause: "The deployment pipeline did not validate the environment variables against the expected schema before promoting to production."
- Contributing factors: "The staging environment does not use the same secrets management as production, masking the error during testing."
- Action items: "1) Add env var validation step to the pipeline. 2) Align staging secrets management with production. 3) Update the deployment runbook."
- Follow-up: "All action items assigned with owners and due dates. Progress to be reviewed in the next SRE sync."
- Lessons learned: "Automated rollback triggers would have reduced MTTR by an estimated 18 minutes in this incident."
Najczęstsze błędy Polaków
1. "deploy" vs "deployment" — czasownik vs rzeczownik. Deploy to czasownik (we deploy the service), deployment to rzeczownik (the deployment failed). ❌ "The deploy failed on staging." → ✅ "The deployment failed on staging." / ✅ "We failed to deploy to staging."
2. "the pipeline is broken" — zbyt ogólne. Precyzyjniejsza wersja określa, na którym etapie. ❌ "The pipeline is broken." → ✅ "The pipeline failed at the integration test stage." albo "The pipeline is failing on the build step."
3. "I will restart pod" — brakuje rodzajnika. Kubernetes pod to policzalny rzeczownik — wymaga rodzajnika. ❌ "I will restart pod." → ✅ "I'll restart the pod." lub "I'll restart the affected pods."
4. "rollback" (rzeczownik) vs "roll back" (czasownik) — dwa słowa vs jedno. Jako rzeczownik lub przymiotnik — jedno słowo: a rollback plan, initiate a rollback. Jako czasownik — dwa słowa: we need to roll back, I'm rolling back now. ❌ "We need to rollback the release." → ✅ "We need to roll back the release."
5. Wymowa "environment" — zagubiony dźwięk. Polacy często mówią "enviroment" — opuszczając drugie 'n'. Poprawna wymowa: en-VY-ron-ment (4 sylaby). W rozmowach technicznych skrót "env" jest powszechnie akceptowany: "Check the env vars."
Quick Reference Table — 28 terminów
| EN Term | PL Tłumaczenie | Typowy kontekst |
|---|---|---|
| pipeline | potok CI/CD | CI/CD |
| build | kompilacja | CI/CD |
| artifact | artefakt | CI/CD |
| deployment | wdrożenie | CI/CD |
| rollback | wycofanie wdrożenia | CI/CD |
| staging environment | środowisko staging | CI/CD |
| production environment | środowisko produkcyjne | CI/CD |
| release | wydanie | CI/CD |
| container | kontener | Infrastruktura |
| image | obraz kontenerowy | Infrastruktura |
| pod | pod | Kubernetes |
| cluster | klaster | Kubernetes |
| node | węzeł klastra | Kubernetes |
| namespace | przestrzeń nazw | Kubernetes |
| helm chart | pakiet Kubernetes | Kubernetes |
| service mesh | service mesh | Infrastruktura |
| uptime | dostępność systemu | Monitoring |
| SLA | umowa o poziomie usług | Monitoring |
| SLO | cel poziomu usług | Monitoring |
| SLI | wskaźnik poziomu usług | Monitoring |
| alert | alert | Monitoring |
| on-call rotation | dyżur on-call | Monitoring |
| postmortem | analiza incydentu | Monitoring |
| MTTR | średni czas naprawy | Monitoring |
| IaC | infrastruktura jako kod | Automatyzacja |
| Terraform | Terraform | Automatyzacja |
| Ansible | Ansible | Automatyzacja |
| idempotent | idempotentny | Automatyzacja |
| drift | dryfowanie konfiguracji | Automatyzacja |
| provisioning | inicjowanie zasobów | Automatyzacja |
Podsumowanie
Angielski w DevOps to nie opcja — to część pracy. Incident response, postmortem, dyskusja z dev teamem o failed pipeline — to wszystko dzieje się po angielsku w każdym globalnym zespole. Mając te 28 terminów w głowie, możesz działać pewnie niezależnie od tego, czy to środek nocy czy standup z Londynem.
Rozszerz swoje słownictwo w artykule o słownictwie IT po angielsku, sprawdź angielski dla Cloud Architekta jeśli pracujesz z infrastrukturą chmurową, lub zajrzyj do angielski dla AI/ML Engineera jeśli Twoje pipeline'y dotykają machine learningu.
Gotowe fiszki z terminologią DevOps znajdziesz w ścieżce DevOps Engineer w sekcji IT & Programowanie.