Empowering Developers with a Database DevOps Strategy

Empowering Developers with a Database DevOps Strategy: Building Guardrails for Speed and Stability

Modern DevOps practices have revolutionized application delivery—but when it comes to stateful systems like PostgreSQL, the story is different. Self-service infrastructure sounds simple until you add databases. Schema changes, configuration conflicts, and replica mismanagement can derail automation and impact production stability.
This presentation from PGConf NYC 2025, led by Phil Vacca, Director and Global Practice Leader for PostgreSQL Services at Datavail, provides a practical roadmap for building a Database DevOps strategy that works. It’s designed to help platform engineering teams reduce friction for developers while maintaining the guardrails that protect critical data.
Key Highlights:

Why Platform Engineering Exists
Understand the role of platform teams in reducing developer friction and why databases require unique strategies beyond standard DevOps tooling.

Postgres-Specific Challenges
Learn why PostgreSQL doesn’t behave like application code. Explore issues like configuration inheritance, replica deployment order, and parameter conflicts that complicate automation.

Building Effective Guardrails
Discover how to enable self-service without chaos. Implement automated workflows with built-in warnings and controls to prevent costly incidents.

What Actually Works
Gain insights from 20+ years of engineering experience, including:

Organizational buy-in tactics.
Success metrics that matter.
Iterative improvements and feedback loops to keep platforms aligned with developer needs.

Common Mistakes to Avoid
Avoid pitfalls such as tool selection without metrics, trying to solve everything at once, and creating communication bottlenecks.

Why This Matters
Your developers want speed. Your data demands security, availability, and performance. A well-designed Database DevOps strategy reconciles these competing priorities—empowering teams to move fast without sacrificing reliability.
Whether you’re running PostgreSQL on-premises, in the cloud, or in a hybrid environment, this session delivers actionable strategies for:

Reducing deployment friction.
Protecting production data.
Measuring success effectively.

About the Presenter
Phil Vacca brings over two decades of experience in engineering and operations, specializing in building scalable, reliable database systems that support developer productivity without compromising stability.

👉 Download the full presentation and learn more here:
https://www.datavail.com/resources/empowering-developers-database-devops-strategy/

1. Empowering
Developers
With a Database DevOps
Strategy

2. What We’ll Talk About
Defining the Terms
Rise of DevOps
Platform Engineering
Platform Engineering Challenges
Measuring the Right Things
Stateful Applications
Postgres Example
Building a Strategy
What Works?
Common Pitfalls

3. www.datavail.com 3
20+ years of Engineering and Operations
team experience
Past: Leading Platform Infrastructure
teams
Currently Global Practice Leader at
Datavail for PostgreSQL services
Extensive expertise in cloud
environments
P
Phil
Vacca
Director & Global
Practice Leader,
@ Datavail
Connect or follow me on LinkedIn:
https://www.linkedin.com/in/
philvacca/

4. www.datavail.com 4
Fill out your session evaluation
form for a chance to win a
LEGO set.
Put the Postgres
Pieces Together

5. www.datavail.com 5
Datavail at
a Glance
Delivering a superior
approach to leveraging
data through application of
a tech-enabled global
delivery model & deep
specialization in databases,
data management, and
application services.
$25
M
Invested
in IP that improves the
service experience and
drives efficiency
18
+
Years
building and operating
mission critical data
and application systems
www.datavail.com 5
Global Team
1,300 Employees
staffed 24x7, resolving over
2,000,000 incidents per year

6. www.datavail.com 6
Convergence of Developer Best Practices and Operational Needs
•Infrastructure-as-Code in Source Control
•CI/CD Pipelines
•Metrics, Monitoring, Observability
Focused on Shipping Product
•Application code in containers
•Cloud hosted services
Reduce Friction To Deploy Features
•Value Stream Mapping
Eliminate Silos
DevOps is a Philosophy, Not a Job
The Rise of DevOps

7. www.datavail.com 7
From Philosophy to Results
Answers the question: How Do We Remove Friction For Developers?
Platform Teams are cross functional
Focus on Developer Success Metrics
•Time to release features and fixes
•Time to deliver new infrastructure
Goal is Developer Self-Service
•Internal Developer Portal (IDP)
Build Guardrails
•Developers want to move fast
•Don’t let them make bad (deployment) choices
Platform Engineering Emerges

8. www.datavail.com 8
What About Stateful Data?
Provisioning Databases for Developers Seems Easy
Sizing and Tuning them for Production is Hard
•RDBMS is traditionally vertical scaled
•Traditional data stores do not auto-scale gracefully
•Serverless is a misnomer, and can be expensive
•Data replicas can increase spend substantially
•Replica queries can feedback to the Primary
Deployment Challenges
•Schema Changes Block
•Settings Changes Can Require Restart

9. www.datavail.com 9
The Postgres Example
Configuration and OS Parameters
•Abstract your configuration files
•Let the Postgres project handle specifics and defaults
Deploy to the Secondary first. . .
•Except many settings must match the Primary
•max worker processes, shared buffers, etc.
ALTER SYSTEM / configuration drift
•You can’t disable ALTER SYSTEM, you can only limit access

10. www.datavail.com 10
What Works
Organizational Buy-In
•Find an executive champion
Have Success Metrics
•And check them regularly
Iterative Approach
•Improve existing pipelines
Focus On Automating Happy Path
•Add clear warnings when something strays
•Automate the boilerplate – ticket creation, pull requests, etc.
Keep Talking to Your Developers
•Build a feedback loop
Build Guardrails
•Watch your spend
•Review new infrastructure / build infra dashboards

11. www.datavail.com 11
Common Pitfalls
Focusing on Tools (Over Metrics)
•A Software Catalog is what we need!
•A Developer Platform is what we need!
Trying to Solve All the Problems
Inventing New Communication Patterns
Overloading The Team
•What is the mission of the Platform Engineering Team?
Not Knowing the Options When You Provide A
Template
•Limit the options / avoid all things to all people

12. Conclusions
DevOps is everywhere
Platform Teams Exist to Serve Developer
Needs
Building Self-Service is Hard
Short Feedback Loops Are Essential
Self-Service is Even Harder When Stateful
Applications Are Involved
Build Guardrails

13. www.datavail.com 13
Fill out your session evaluation
form for a chance to win a
LEGO set.
Put the Postgres
Pieces Together

14. www.datavail.com 14
Thank You
datavail.com
[email protected]
866-811-6764