What’s the Best PostgreSQL High Availability Framework? PAF vs. repmgr vs. Patroni Infographic

Wondering which high availability framework to use for your PostgreSQL deployments? We compared the top 3 frameworks, PostgreSQL Automatic Failover (PAF) vs. Replication Manager (repmgr) vs. Patroni, in our Managing High Availability in PostgreSQL series. In these blogs, we dove deep into how the frameworks work, their setup requirements, pros and cons, and how they performed in standby server tests, primary server tests and network isolation tests (split brain scenario) to help you determine the best framework to improve the uptime for your PostgreSQL-powered applications.

PostgreSQL High Availability Framework Infographic

Given the popularity of these posts, we condensed the results into the below infographic so you can easily visualize and compare how the top frameworks handle these common failure scenarios.

Download the infographic

What's the Best #PostgreSQL High Availability Framework? PAF vs. repmgr vs. Patroni InfographicClick To Tweet

So, which PostgreSQL high availability framework are you leveraging for your application? We’d love to hear how it’s working for you, and how you’d like to improve the performance on your deployment. Let us know your thoughts in our comments, or send us a message on Twitter at @scalegridio!

Download the infographic

If you need help setting up high availability for your PostgreSQL clusters, check out our fully Managed PostgreSQL on AWS, PostgreSQL on Azure, and PostgreSQL Enterprise solutions to automate your database management in the cloud an on-premise. We work with thousands of DBA’s, DevOps engineers, and developers to improve their availability, performance and security, and would love to help you out as well. Explore the platform in a free 30-day trial - no credit card required, or visit our Compare PostgreSQL Providers page to see why we’re the #1 PostgreSQL hosting solution.

Embed This Infographic On Your Site

Embed This Image On Your Site (copy code below):

Kristi is the Marketing Manager at ScaleGrid. She has over 10 years of experience working with high-tech startups, with a background in marketing, finance, operations, and product management.

  • ioguix


    There’s three entries I don’t understand:

    * “primary server: Kill the PostgreSQL Process“: both PAF and Patroni have a green check but one is “Downtime” the other one is “No downtime”. Is it a typo? If not, what’s the difference?

    * “primary server: Reboot the Server“: both PAF and Patroni show “Downtime”, but PAF is orange and Patroni is green checked. What’s the difference there? Is it the auto-pg_rewind on the Patroni side?

    * “Stop the framework agent process“: I’m not sure to understand the point of this test. Is it to simulate the impact of a graceful stop of the framework, eg for cold upgrade?

    • Madan Kumar


      * “primary server: Kill the PostgreSQL Process”: Yes, there were typos in the infographic and thanks for pointing it out. We have updated the infographic. So there is downtime in all frameworks for this case.

      * “primary server: Reboot the Server”: Yes, auto pg_rewind(parameter must be set to true) is the advantage in case of Patroni. In PAF, there is no auto rewind and it will require manual intervention if there is a timeline divergence.

      * “Stop the framework agent process”: This test case was to determine the behavior when the framework agent is crashed or killed due to memory crunch. The graceful stop may not depict the actual scenario but it was helpful to figure out the implication that is possible when a framework agent is down. This will convince the need for use of watchdog.