2. Market Leading
Data Security
Crunchy Data is the leader in
PostgreSQL security. Common
Criteria certification and essential
security enhancements make Crunchy
Certified PostgreSQL the trusted open
source PostgreSQL distribution for
the enterprise.
Cloud Ready
Data Management
Whether deploying to public or
private clouds, Crunchy Data
provides market leading, open
source, Kubernetes-based
technology solutions, giving your
team the choice and flexibility for how
you deploy your data.
Leader in Open Source
Enterprise PostgreSQL
Crunchy Data gives organizations the
technology, support, and confidence to
enjoy the power and efficiency of open
source PostgreSQL.
3. 3
• Vice President of Platform Engineering, Crunchy Data
• Previously: Engineering leadership in startups
• PostgreSQL Major Contributor
• Advocacy & various committees for PostgreSQL
Global Development Group
• @postgresql + .org content
• Director, PgUS
• Conference organization + speaking
• @jkatz05
About Me
5. • The PostgreSQL relational database is over 30 years old and has been open
source since 1996
• Open source like Linux: no single vendor; flexible license; no one can own
PostgreSQL
• Recent PostgreSQL releases have brought it to feature parity with popular
proprietary databases
• PostgreSQL's extensibility allows it to handle NoSQL workloads,
accommodate developer language preferences, and provide full transaction
safety and data integrity
• PostgreSQL’s stability, consistency, and robustness have made it the trusted
open source database in the enterprise.
PostgreSQL: An Oldie Stays a Goodie
6. • Stateful workloads have only one job: maintain state
• If you lose or corrupt all your data, you're done
• Specific knowledge is required about a stateful application (e.g.
PostgreSQL) to perform state modification operations such as provisioning,
failover, or recovery. A Kubernetes Operator can do this.
• Steps to do state modification operations range from simple to tedious
• Adding a user to a database is simple
• Adding a user to 1,000 databases is tedious
• Proper Operator design allows for autonomous managed workloads:
distributed consensus HA, systematic backups, etc.
The Need for Kubernetes Operators
7. • Open Source! GA March 2017, Apache 2.0 Licensed
• https://github.com/CrunchyData/postgres-operator
• Current version: 4.4.1; 4.5.0 beta 1
• OperatorHub.io: Level 5 “Auto Pilot” Capabilities
• Supports essential database-as-a-service functionality:
• Provisioning (create, delete clusters, clone existing clusters)
• High-availability (distributed consensus based, leverages pod anti-affinity,
supports synchronous replication)
• Elasticity (add/remove replicas)
• Disaster recovery (backup, restore geared towards terabyte scale DBs,
automatically schedule backups)
• Administration (PostgreSQL software updates, user management)
• Connection pooling via pgBouncer
• Monitoring via pgMonitor
Crunchy PostgreSQL Operator
8. • Automation: Complex, multi-step tasks reduced to one-line commands
• Standardization: Many customizations, same workflow
• Ease-of-Use: Simple API. Can add on with CLI, UI
• Scale
• Provision & manage quickly amongst thousands of instances
• High-Availability, Load balancing, disaster recovery, security policies,
deployment specifications
• Security: Sandboxed environments, RBAC, mass grant/revoke policies
• Flexibility: Run your stateful workload in any environment you have a node
Why Use an Operator?
16. Crunchy PostgreSQL Operator:
Production Deployments from a Single Command
pgo create cluster hacluster --metrics --pgbadger –pgbouncer --pgbackrest-storage-type="local,s3"