Dash Enterprise 6.2.0

The Dash Enterprise 6.2.0 release focuses on three things:

• Resource management
• Reliability
• Security

Each is a step toward a platform that asks less of your cluster, holds up better under pressure, and fits the security posture of the environments it runs in.

Resource management

The core platform's default footprint is down by roughly half, in both memory and CPU. We measured what the platform's own workloads use under load and set each reservation to match, so Dash Enterprise 6 needs less of your cluster to run and leaves more of it for the applications you deploy.

Those values are no longer fixed, either. Every CPU and memory request and limit — for the platform's own services and for your apps, workspaces, and managed databases — is configurable, and app memory limits can now go above the previous 24 GB ceiling. You size the platform to your cluster rather than the other way around.

Reliability

6.2.0 improves how the platform behaves under pressure, in two ways.

The first is resource behavior. With the hardcoded CPU limits removed, workloads aren't throttled when they need CPU, so they run faster and more predictably under load. Memory is sized to each core workload's real usage plus headroom, so a service that a too-tight default used to OOM-kill now has the room it needs; and because its memory request equals its limit, it stays within its reservation and isn't first in line for eviction when a node runs short. The same resource work that shrinks the footprint makes the platform steadier.

The second is recovery. Backup and restore now run on a durable execution engine: a backup or restore interrupted by a pod restart or a node failure resumes and completes instead of failing partway through, so the recovery you depend on is there on the day you need it.

Security

6.2.0 continues to harden Dash Enterprise 6 for the locked-down clusters enterprises run it in. Every platform and app workload now runs on a read-only root filesystem — apps that legitimately need to write to disk, such as from a predeploy or postdeploy script, can be given a writable path in the Helm chart. Seccomp profiles are applied at the container level and now cover Dash app, workspace, and service workloads. And there are stronger safeguards against sensitive data making its way into logs.

New to Dash Enterprise? Request a demo to see it running against a cluster like yours.