Database Support Engineer at Percona. 24x7x365 support for large enterprise clients across multiple industries. Hundreds of requests resolved where MySQL and PostgreSQL were pushed to their limits under tight SLAs.

The Context #

Enterprise databases under pressure – replication breaking, queries degrading, clusters failing. Large companies with complex infrastructure where downtime meant direct revenue loss. The job was to get in, diagnose, and fix production systems within the SLA.

The Work #

Systematic troubleshooting under pressure. Reproduce, isolate, fix, document. Every request became a runbook for the next one. Diagnosed complex problems using Percona Monitoring and Management (PMM), slow query analysis, and deep knowledge of MySQL and PostgreSQL internals.

Performance optimization. Query optimization, indexing strategies, and scaling architecture on AWS (Aurora, RDS) and GCP.

Knowledge transfer. Trained engineering teams on database best practices. Created technical documentation, knowledge base articles, and operational runbooks. Active in open source community – bug reporting and confirmation. Built automation scripts that saved time across client environments.

MySQL (5.6–8.0), PostgreSQL, MongoDB (3.x–5.x), AWS Aurora/RDS, GCP, Docker, PMM, Prometheus, Python/Bash automation.

Lessons Learned #

• Discipline under pressure scales. A consistent method (reproduce, isolate, fix, document) works whether the database is 10 GB or 10 TB.
• Query optimization compounds. Well-tuned queries outperform hardware upgrades every time.
• Good documentation saves everyone time. Clear runbooks mean the next person solves it in minutes, not hours.
• Reproduce it yourself. The fastest way to understand a problem – or a tool – is to reproduce the behavior with your own hands. Configuration edge cases, replication quirks, failure modes. Reading about it is not knowing it.