
Infrastructure security: an audit first, then fixes in order of risk
We check who has access to what, where secrets are kept, what is reachable from the internet and whether backups can be restored. You get a list of findings ordered by risk. Then we close them through code, so the fixes stay in place.
When you need this
- 1
A client or a partner asks how your infrastructure is protected
- 2
Keys and passwords sit in the repository or in a shared chat
- 3
Everyone on the team has admin rights in the cloud account
- 4
A database or an admin panel is reachable from the internet
- 5
Backups are made, but nobody has ever restored from one
What's included
Audit
Cloud accounts, servers, clusters, CI and repositories are checked for excessive access, open ports, leaked secrets and outdated software. The result is a report with risks in order of severity.
Access: IAM and RBAC
Personal accounts instead of shared ones, roles with the minimum permissions, MFA, no long-lived keys where short-lived credentials work. The same in Kubernetes through RBAC.
Secrets
Secrets are removed from code and from plain-text CI variables. They are encrypted with SOPS in Git or kept in Vault or a KMS-backed cloud store. Keys that have leaked are rotated.
Network
Databases and internal services move to private networks: private subnets of a VPC on AWS, a private network elsewhere. Only the load balancer faces the internet, and administration goes through a VPN. A WAF and rate limits stand in front of public endpoints.
Backups with tested restores
Backups are encrypted and stored outside the server and the account they protect. We run a restore and record how long it takes.
Patching
A process for OS, Kubernetes and image updates: automatic security updates where that is safe, scheduled windows for the rest, image scanning in CI.
How the work goes
- 1
Call and access
You describe the system and what worries you. We get read-only access for the audit.
- 2
Report and price
Findings with a severity and a proposed fix for each. You choose what to fix, and the price is agreed before the work.
- 3
Hardening
Fixes go through pull requests, the riskiest first. Changes that could lock people out are agreed and scheduled in advance.
- 4
Handover or support
You get the access model, the rules for handling secrets and the restore procedure in writing. Under support we keep patching and review access regularly.
Questions
- AWS IAM
- Kubernetes RBAC
- SOPS
- Vault
- AWS KMS
- AWS VPC
- WireGuard
- Cloudflare
- Trivy
- Terraform
- Hetzner
Describe the task — get a plan and a price
Write what you have and what you need. We answer within one business day with questions or a concrete proposal.