Cloud + DevOps
We architect, build, and operate cloud infrastructure for SaaS, e-commerce, and internal-tools clients. Our own infrastructure (5 SaaS products, ~7 servers, 5 RDS databases, ElastiCache, CloudFront) is the proof of practice — the same engineers who run our production run yours.
Konkrete Ergebnisse
- AWS / GCP / Azure architecture from scratch — VPC, IAM, networking, secrets management.
- Kubernetes clusters — EKS, GKE, AKS, or self-managed (K3s for small fleets).
- CI/CD pipelines — GitHub Actions, GitLab CI, CircleCI, custom Jenkins migrations.
- Infrastructure as code — Terraform, Pulumi, AWS CDK, Cloudformation. State management + module libraries.
- Observability — Datadog, New Relic, Grafana + Prometheus + Loki, OpenTelemetry instrumentation.
- Security + compliance — SOC 2 prep, GDPR/HIPAA controls, secrets rotation, IAM audits.
Womit wir arbeiten
Wir wählen, was zu Ihrem Team passt — keine erzwungenen Präferenzen.
Cloud providers
AWS (deepest) · GCP · Azure (less common) · Hetzner (cost-optimised EU)
Container orchestration
Kubernetes (EKS/GKE/AKS) · ECS Fargate · Docker Swarm (small ops) · Nomad (rare)
IaC
Terraform · Pulumi · AWS CDK · Helm · Kustomize
CI/CD
GitHub Actions (default) · GitLab CI · CircleCI · Jenkins (legacy migrations) · ArgoCD for GitOps
Observability
Datadog · New Relic · Grafana + Prometheus + Loki · Sentry · OpenTelemetry
Databases
RDS (PostgreSQL, MySQL) · Aurora · DynamoDB · Redis (ElastiCache) · ClickHouse · Snowflake
So arbeiten wir
Discovery (1-2 weeks)
Read your existing infra (if any), document the current state, write a target-state architecture diagram. Risk + cost matrix in writing.
Foundations (2-4 weeks)
VPC, IAM, secrets, IaC scaffolding, CI/CD pipeline. Nothing in production yet — but the rails are laid.
Migrate / cutover
Phased migration with rollback plans. Often we run old + new in parallel for 1-2 weeks.
On-call + observability
PagerDuty / OpsGenie integration, runbooks, observability dashboards, optional 24/7 on-call retainer.
Aus unserer eigenen Produktion
Merot platform (5 SaaS products)
Single EC2 t3.large + shared RDS + ElastiCache + CloudFront serving 5 production SaaS apps + marketing sites.
Lambda@Edge subdomain redirects
CloudFront Function deployed across 6 distributions to 301 marketing paths from per-product subdomains to merot.com/{product}/* — single source of truth in git.
PDF generation infra
Node + Handlebars + WeasyPrint running on a single VPS, serving 3 SaaS products via shared template engine — survives spikes via CDN-cached PDFs.
Engagement-Modell
Häufige Fragen — Cloud + DevOps
Do you work with our existing cloud account, or do you provision a new one?
Yours. We work as an external IAM principal (often via SSO + a dedicated 'merot-engineer' role). We don't take ownership of your AWS account or your billing.
What about SOC 2 / ISO 27001 / HIPAA prep?
Yes — we've taken multiple clients through SOC 2 Type II audits. We'll help you scope what's in your trust boundary, write the security policies, set up the technical controls (access logging, MFA, secrets rotation), and prep for the auditor's evidence requests.
How do you handle production secrets?
AWS Secrets Manager / GCP Secret Manager / Vault — never .env files committed to git. CI/CD reads from the secret store at deploy time. We help you rotate after our engagement ends.
Kubernetes — yes or no?
Default no for teams under 10 services. Often ECS Fargate or even just systemd on a VM is cheaper + simpler. We push back on K8s when the operational cost outweighs the benefit.
Multi-region / global distribution?
Yes — including active-active Postgres setups, CloudFront geo-routing, RDS read-replicas across regions. Heavier engagement (3-6 months typical).
Cost optimisation engagements?
Common. We audit your AWS bill, find the cuts (rightsizing, savings plans, spot instances, CloudFront cache-hit improvements, RDS instance type changes), and implement. Typical 30-60% reduction in 1-2 quarters.
On-call rotation — do you provide it?
Optional. We can be your tier-2 on-call (24/7 with paging via PagerDuty / OpsGenie) on retainer. Typical SLA: 15-min ack, 30-min triage, 1-hour mitigation for P1.
Do you migrate from one cloud to another?
Yes — common engagement. AWS → GCP, Azure → AWS, on-prem → cloud. We scope, plan rollback, run a phased migration. Plan 3-6 months for a full SaaS migration.
Cloud + DevOps-Projekt scopen
60 Min. Discovery-Call kostenlos. 6-seitiger Plan in 48 h.