Stabilize & recover
Production Linux troubleshooting, inherited infrastructure, failed upgrades, migrations, storage and networking issues—handled methodically, with the evidence preserved.
Linux, cloud & infrastructure consulting · Constanța / Europe
Infrastructure
without the drama.
Senior Linux and cloud operations help for small SaaS teams, hosting companies and agencies—without adding a full-time hire.
20+years in production infrastructure
LinuxOpenStackCloudStackCephAnsibleElastic
How I can help
Senior operations,
sized for the problem.
Focused engagements for teams that need experienced infrastructure judgment, not another layer of process.
See problems sooner
Logging, dashboards and alerts that answer useful questions. Backup and restore reviews that verify recovery instead of merely reporting a green checkmark.
Automate without surprises
Repeatable Ansible and Python workflows, careful rollout plans, and hardware automation for environments where reliability matters more than novelty.
Selected work
Experience you can
put to work.
Examples from hands-on infrastructure operations. Client identities and internal system details are omitted.
01
Making cloud infrastructure logs searchable
- The problem
- Investigating incidents meant piecing together logs spread across cloud infrastructure nodes and services.
- My contribution
- Built an ELK collection and search workflow, with structured log fields, Kibana views, deployment automation and retention controls. Added checks for collection health and archive capacity.
- The outcome
- A shared place to investigate service events, supported by repeatable deployment and visibility into the collectors themselves.
02
Turning maintenance checks into useful decisions
- The problem
- Archive usage kept rising despite age-based retention. A single disk percentage could not explain whether cleanup was broken or simply waiting for its next run.
- My contribution
- Developed collector reporting and checked filesystem usage alongside timer state, service results, expired files and protected archives. Distinguished historical resource spikes from current pressure.
- The outcome
- Confirmed working retention and identified the capacity guard pausing new archive copies. Documented the evidence and options for cleanup cadence and storage growth.
03
Isolating a multi-subnet load-balancer failure
- The problem
- New load balancers accepted frontend connections but could not reach healthy backend members in another subnet.
- My contribution
- Compared working and failing instances, checked member subnet assignments and available addresses, and reproduced the behavior after a supported failover.
- The outcome
- Isolated incomplete member-subnet attachment and prepared a focused vendor escalation with a preserved reproducer. The investigation established the fault boundary; a vendor repair was still pending at handoff.
ADRIAN AVADANI · CONSTANȚA
About Adrian
Twenty years of
keeping production honest.
I’m Adrian Avadani, a Linux and cloud infrastructure engineer based in Constanța, Romania. For more than two decades, I’ve operated hosting platforms, production systems and private-cloud environments—often the mature, business-critical kind that cannot simply be rebuilt from scratch.
My background spans hands-on Linux operations, OpenStack and distributed storage, incident response, observability, automation and bare-metal infrastructure. I work best where technical depth, calm judgment and clear communication matter equally.
I don’t arrive with a preferred rewrite. I start by understanding what exists, what is actually failing and what the business needs the system to do.
20+ yearsProduction infrastructure
Linux to cloudBare metal, virtualization, OpenStack
Constanța, RomaniaRemote across Europe
How I work
Calm is
a technical skill.
Production incidents reward discipline. Mature systems reward context. Both punish guesswork.
01
Evidence before action
Find the failing layer and establish a live source of truth before changing production.
02
Small, reversible changes
Pilot narrowly, verify the result, then expand. Every change needs a way back.
03
Leave the system clearer
Useful documentation and runbooks are part of the fix, not an optional deliverable.
A useful first engagement
Start with an
infrastructure review.
I’ll map the important moving parts, review monitoring and recovery readiness, and identify the few risks actually worth addressing first.
- Production risk and reliability review
- Monitoring, logging and alerting gaps
- Backup and restore readiness
- Prioritized, practical next steps
What you receive
- A map of the key systems and dependencies
- Findings ranked by impact, with evidence and unknowns
- An action plan covering next steps and rollback considerations
- A walkthrough of the findings and recommendations
Discuss the situation
Tell me what runs, what concerns you and what a useful outcome looks like.
Agree the scope
We agree the systems, access, deliverables, timeline and fee before work begins.
Review, then decide
Start with observation and documentation. Any production changes are agreed separately.
Suitable for an inherited Linux environment, recurring incidents or a review before a migration. Coverage depends on the agreed scope and available access; this is not a 24/7 emergency-response service.
Discuss an infrastructure review ↓