In the world of cloud databases, "High Availability" (HA) and “Disaster Recovery” (DR) are often treated as checkboxes. But not all HA/ DR configurations are engineered to survive under the stress of production enterprise workloads.

For mission-critical applications, the gap between standard cloud recovery and true database resilience is measured in dropped connection pools, missing transactions, and lost session states.

Join Jags Ramnarayan, CTO Cloud Innovations at MariaDB on May 28th for an architectural exploration on how database failures impacts business continuity, and a deep dive into the technical limits of standard Hyperscaler DBaaS HA setups:
  • The Business Impact of Failover: Real world look at how standard cloud recovery windows disrupt application-tier workflows (and how InnQuest resolved these limits).
  • The DNS Propagation Bottleneck: Why relying on DNS TTL updates to route traffic to a standby node guarantees a 60-to-120-second application backout.
  • The Standby Recovery Trap: Why passive standby nodes lead to time-consuming crash recovery before they can safely accept write traffic.
  • Session State Amnesia: How standard, nontransparent proxies drop prepared statements and active transactions during a failover shift.
We will bypass the high-level vendor promises and put these architectures to the test. Through a live, side-by-side failover simulation, we will demonstrate how to transition from brittle active-passive standbys to a state-preserving, active-active clustering architecture. You will see firsthand how a modern data layer preserves application state, automatically replays in-flight transactions, and keeps your services online without throwing a single client-side error.

Register Now

SPEAKER AND HOST

Speaker 1

Jags Ramnarayan

Speaker
CTO Cloud Innovations
MariaDB

Speaker 2

Fermi Fang

Host
Senior Product Marketer
MariaDB