Skip to main content

Database Maintenance

Your data is the one thing you cannot reinstall

Applications can be redeployed; lost or corrupted data often cannot be rebuilt. We maintain the databases behind your applications so backups restore, replicas keep up and upgrades happen before support ends.

Queries optimisedBackup verified

Business impact

Database neglect turns into data loss or downtime

Databases tend to work quietly until they reach a limit: a full disk, a replica that has silently stopped, a backup that cannot be restored. These are among the most disruptive failures a business can face, and most are preventable with routine care.

Untested backups are an assumption

A backup job that reports success is not proof of a usable backup. Only a restore test confirms that data, permissions and consistency are intact.

Growth slowly degrades performance

Queries that ran comfortably on a small dataset can become slow as tables grow, especially when indexes are missing or no longer match how the data is queried.

Database versions reach end of life too

MySQL, PostgreSQL, MariaDB and MongoDB releases all have published support lifecycles. Running past them leaves known issues unpatched.

Common issues

Problems we regularly solve

Backups that will not restore

Dumps that are incomplete, inconsistent across tables, missing routines or users, or stored on the same server as the database itself.

Replication silently broken or lagging

MySQL or MariaDB replicas stopped on an error, or PostgreSQL standbys falling behind, leaving failover targets out of date.

Disk space exhausted

Binary logs, write-ahead logs or table bloat filling the disk until the database stops accepting writes.

Missing or redundant indexes

Full table scans on busy queries, alongside unused indexes that slow every insert and update.

PostgreSQL vacuum falling behind

Autovacuum unable to keep up with heavy updates, causing table bloat and, if left long enough, transaction ID wraparound risk.

Over-privileged database accounts

Applications connecting as root or superuser, shared credentials across environments and databases reachable from the public internet.

What's covered

What's included in database maintenance

Database maintenance focuses on durability, availability and health. Query and configuration tuning for speed is covered in depth by database optimisation.

Backup and restore verification

Consistent logical and physical backups, point-in-time recovery where supported, off-site copies and scheduled restore tests with recorded results.

Replication and high availability

Monitoring of replica status and lag, failover readiness checks and documented recovery procedures for primary failure.

Index and table health

Review of slow query logs and index usage, removal of unused indexes, and vacuum, analyse or optimise routines appropriate to each engine.

Version upgrades

Planned major and minor version upgrades, tested against a copy of production data with application compatibility checks before cut-over.

Capacity monitoring

Tracking disk usage, connection counts, log growth and table sizes so limits are addressed before they cause an outage.

Access and security

Least-privilege accounts per application, encrypted connections, network restrictions and regular credential rotation.

Maintenance activities

What we do, and how often

Exact frequencies depend on your plan and platform. This is a typical schedule.

Daily

  • Verify backup jobs completed and copies reached off-site storage
  • Check replication status and lag
  • Monitor disk usage and connection saturation

Weekly

  • Review slow query logs and error logs
  • Check vacuum and analyse activity on PostgreSQL

Monthly

  • Restore a backup to an isolated environment and validate the data
  • Review index usage and table growth trends
  • Apply minor version and security updates via staging
  • Maintenance report with backup, capacity and health findings

Quarterly

  • Review database user privileges and rotate credentials
  • Check engine support lifecycles and plan major upgrades
  • Rehearse documented failover or recovery procedure

Our methodology

How we work

01

Assess the estate

We record each database's engine, version, size, hosting, backup method and replication set-up, and test a restore before changing anything.

02

Secure the backups first

Backup coverage, retention and off-site storage are corrected so that every later change has a verified safety net.

03

Stabilise and monitor

Monitoring for disk, replication, connections and slow queries is put in place, and immediate risks such as full disks or broken replicas are fixed.

04

Maintain and upgrade

Routine health work runs on schedule, and version upgrades are rehearsed on production copies before they are carried out live.

Business benefits

What you gain

Backups you have actually tested

Regular restore tests give evidence, not hope, that your data can be recovered.

Earlier warning of capacity limits

Trend monitoring is designed to flag disk, connection and growth issues while there is still time to act calmly.

Upgrades without guesswork

Rehearsing upgrades on copies of real data helps reduce surprises during the live change.

Tighter access to sensitive data

Least-privilege accounts and network restrictions reduce the impact if an application credential is ever exposed.

Documented recovery steps

Written, rehearsed procedures mean recovery does not depend on one person remembering what to do.

Technologies supported

  • MySQL
  • MariaDB
  • PostgreSQL
  • MongoDB
  • Redis
  • Amazon RDS and Aurora
  • Google Cloud SQL
  • Azure Database
  • Percona XtraBackup
  • mysqldump and mydumper
  • pg_dump and pgBackRest
  • MongoDB Atlas
  • ProxySQL and PgBouncer
  • Prometheus exporters

Running something not listed? Ask us – we'll tell you honestly if it's a fit.

Who it's for

Who needs database maintenance?

Businesses running self-hosted databases

Organisations with MySQL, PostgreSQL or MongoDB on their own servers or VPS and no dedicated database administrator.

Teams on managed cloud databases

Companies using RDS, Cloud SQL or Atlas who still need restore testing, index care, access reviews and upgrade planning.

Data-sensitive organisations

Healthcare, finance and education organisations where data integrity and recoverability are under scrutiny.

Industry guidance: Healthcare · Financial Services · SaaS Companies · Education

FAQs

Database Maintenance: common questions

Our cloud provider takes automated backups. Is that enough?
Managed snapshots are a good foundation, but they share the provider account and retention limits with the database they protect. We usually recommend a separate copy, a defined retention policy and regular restore tests so you know how long recovery takes and what it returns.
Will you change our database schema?
Routine maintenance does not change your schema beyond adding or removing indexes, and those changes are agreed with you first. Schema changes needed for new features are development work handled by your developers or scoped separately.
How do you upgrade a database without losing data?
We take and verify a fresh backup, rehearse the upgrade on a copy of production, test the application against it, then schedule the live upgrade with a rollback plan. For larger databases, replication-based approaches can shorten the downtime window.
What access do you need?
An administrative database account for maintenance tasks, read access to monitoring and logs, and access to backup storage. Credentials are stored in a vault, used only by the engineers on your account and can be revoked at any time.
Does this cover data protection compliance?
We operate technical controls such as encryption in transit, access restriction, backup retention and audit records, which support obligations under frameworks like GDPR. We do not provide legal or compliance certification advice.
What if the database fails outside business hours?
Response depends on your plan and the severity levels agreed in your SLA. Because recovery procedures are documented and rehearsed in advance, the steps to restore service are already known when an incident occurs.

Find out whether your backups would really restore

Request a technical audit and we will test a restore, review replication and access, and report on the risks in your database estate.