The rise of DevOps practices is forcing a significant rethink of how databases are managed, deployed, and updated. A global survey of more than 2,000 developers, database specialists, and IT leaders — conducted by Redgate Software, a provider of database management and automation tooling — found that nearly half of respondents (49%) are now deploying database changes to production at least once a week.
That cadence would have been unthinkable in traditional database administration, where change control processes often meant updates were batched monthly or quarterly. The shift reflects a broader cascade: as development and operations teams embrace continuous delivery, the database can no longer remain a slow-moving outlier.
DevOps Adoption Is Growing — But Unevenly
The survey paints a picture of an industry in transition rather than one that has completed it. According to Redgate’s findings:
- 18% of respondents have adopted DevOps across all projects.
- 36% have adopted DevOps across some projects.
- 14% are at the proof-of-concept stage.
- 13% have no plans to adopt DevOps at all.
Adoption is furthest along in financial services and insurance. In those sectors, more than 77% of respondents said their organizations are either in proof-of-concept (16%), have adopted DevOps across some projects (41%), or have adopted it across all projects (20%). This is notable given that regulated industries are often assumed to lag on DevOps adoption due to compliance constraints.
Higher DevOps Maturity Correlates With Fewer Defects
The survey suggests that organizations further along in their DevOps journey see measurable quality improvements. Well over a third (37%) of DevOps adopters report that 1% or fewer of their deployments introduce code defects requiring hotfixes, compared to 30% for groups that have not adopted DevOps practices.
Database deployments specifically show a similar pattern. More than half of respondents (55%) said they perform most or all of their business-critical database deployments while systems remain online — often called zero-downtime or online deployments. Of that group, 75% report that 10% or fewer of those deployments caused defects requiring hotfixes.
Nearly half of all respondents (46%) are also employing some level of automation for database deployments, reflecting growing awareness that manual database change processes are incompatible with continuous delivery pipelines.
Key Barriers to Database DevOps Adoption
Despite the progress, the survey identifies persistent friction points that slow adoption:
- Workflow and business disruption (23%): Integrating database changes into CI/CD pipelines requires changes to established processes, and teams are understandably cautious about disrupting systems that underpin critical business functions.
- Lack of skills (22%): Database DevOps requires practitioners who understand both database internals and modern automation tooling — a combination that is still relatively rare.
- Misalignment between developers and operations (16%): The cultural divide between development and operations teams is well-documented, but the equivalent friction between developers and DBAs is often underappreciated.
Databases Must Be Brought Into the SDLC
Kendra Little, DevOps advocate at Redgate Software, argues that database management needs to be far more tightly integrated with the software development lifecycle (SDLC) as organizations transition to DevOps. The survey data supports this view: over a quarter of respondents (26%) said it is still somewhat or very difficult to get code reviews for database changes. A further 16% said no code reviews are conducted for database changes early in the development cycle.
This is a meaningful gap. In a mature DevOps environment, schema changes, stored procedures, and migration scripts should be version-controlled, peer-reviewed, and tested with the same rigor as application code. When database changes bypass that process, they become a source of unpredictable risk — especially in online deployments where a bad migration can affect live users.
Who Owns the Database? The DBA Role Is Evolving
One of the more nuanced findings concerns ownership of database management itself. The survey shows a fragmented picture:
- 43% of organizations still have a dedicated database administrator (DBA) managing their databases.
- 34% rely primarily on developers.
- 21% use a combination of both.
Little notes that over time most organizations will shift toward a hybrid model, where developers take more ownership of day-to-day database work while DBAs focus on performance tuning, architecture, security, and governance. That transition requires deliberate cultural change. In many engineering organizations, DBAs have historically been perceived as a bottleneck — a gatekeeper who slows down deployments. That perception, while sometimes warranted, often overlooks the legitimate risk management role DBAs play.
The more productive framing is to treat DBAs as platform engineers for the data layer: people who define standards, build automation guardrails, and enable developers to move quickly without inadvertently corrupting production data or degrading query performance.
Practical Takeaways for Teams
- Version-control your database schemas. Treat migration scripts as first-class code artifacts stored alongside application code in source control.
- Add database changes to your CI/CD pipeline. Automate schema validation, testing, and deployment rather than handling them as manual, out-of-band operations.
- Enforce code review for database changes. At minimum, a second set of eyes should review schema migrations before they reach production.
- Plan for online deployments. Design migrations to be backward-compatible where possible, so they can be applied without taking systems offline.
- Reframe the DBA role. Position database specialists as enablers of developer velocity rather than approvers who slow it down.
The broader message from the Redgate survey is clear: database management cannot remain an island isolated from the rest of the DevOps toolchain. As deployment frequency increases and business expectations for uptime harden, organizations that fail to bring their database practices in line with modern SDLC standards will find that their data layer becomes the ceiling on how fast they can actually ship.
Survey data cited in this article is sourced from Redgate Software’s global survey of more than 2,000 developers, database specialists, and IT leaders, as reported by devops.com (original publication date: March 2020).