Database Engineering & Infrastructure
Production MySQL and Percona database audits, performance engineering, high availability, independent monitoring, secure backups and recovery.
Production databases need more than a running MySQL process. We help diagnose performance issues, engineer reliable database infrastructure, detect failures independently and make recovery practical.
MySQL · Percona XtraDB Cluster · ProxySQL · MariaDB · PostgreSQL
Discuss your database → · See pricing →
Database Audit & Performance
- MySQL/Percona configuration, InnoDB, slow queries, indexing, locks and resource usage.
- Replication and cluster health, access control, network exposure and operational risks.
- Clear findings with prioritized, actionable recommendations.
Database Engineering & High Availability
- Troubleshooting, tuning, migrations and replication changes.
- Percona XtraDB Cluster / Galera, ProxySQL, HAProxy and Keepalived.
- Writer/reader routing, quorum, failover behavior, maintenance and recovery runbooks.
External Monitoring & Alerting
Monitoring should survive the failure it is supposed to detect. We favor independent, off-cluster monitoring instead of relying solely on services inside the database environment.
- Prometheus, Alertmanager and Grafana; node and MySQL exporters.
- PXC
wsrepcluster status, node availability, flow control and resource saturation. - Backup freshness and failures, disk space, service health and actionable notifications.
Explore monitoring services → · Read the monitoring engineering article →
Secure Backup & Recovery
A backup is useful only when it is isolated, recent, recoverable and tested.
- Separate backup server pulling data from the database environment.
- Restricted SSH access, forced commands and narrowly scoped sudo permissions.
- Encrypted transfer/storage where appropriate, retention, off-host copies and restore drills.
- External backup-age alerts and documented recovery procedures.
Explore backup & recovery services → · Read the secure PXC backup article →
Production Infrastructure Experience
High-availability database layer for an ISP serving approximately 20,000 subscribers.
An engineering engagement covering a three-node Percona XtraDB Cluster, three ProxySQL instances, Keepalived VIP, writer/reader routing, independent Prometheus/Alertmanager/Grafana monitoring and a separate backup server. Architecture and practices are described without identifying the customer.
Read the PXC / ProxySQL / Keepalived case →
Start With Your Actual Problem
Slow queries? Unstable replication? Cluster failovers? Missing backup alerts? No tested restore procedure?
Send the database version, topology, approximate workload, symptoms and desired outcome. Please do not send passwords, private keys, database dumps or other secrets in an initial inquiry. Work starts after scope and access are agreed.
Packages & pricing
Choose the outcome you need: understand the current state, get a practical implementation plan, or have the agreed changes delivered in production. REVIEW and PLAN do not include production changes.
3–5 business days
- Review configuration, workload and database health
- Identify performance, replication and access risks
- Receive a concise findings report
5–10 business days
- Everything in REVIEW
- Get tuning, upgrade or topology recommendations
- Receive a change plan with validation and recovery steps
Schedule agreed after assessment
- Assessment and implementation plan included
- Execute agreed database changes in production
- Verify results and hand over operating instructions
Indicative delivery times start after the agreed kickoff and receipt of the required access and information. Scope, dependencies and maintenance windows may affect the schedule; timing is confirmed before work.
Each package is a separate engagement; you do not need to purchase all three. Scope and deliverables are agreed before work. Implementation starts from €999; the final quote depends on system count, complexity and the changes required. Large migrations, emergency recovery and ongoing support are quoted separately.
Tell me about your database environment
Describe your environment, the problem and the outcome you need. Scope and access are agreed before work. Do not send passwords, keys, database dumps or other secrets.