Skip to content

v3.0.0

Required Action

Schedule downtime for the upgrade from v2. Stop new checkouts and let active sessions finish. Stop every v2 orchestrator, proxy, and session worker before starting v3. Users must sign in again after the upgrade. Upgrade from Bridge v2 to v3

  • With v2 stopped, back up PostgreSQL and the matching recordings. Keep your configuration, deployment secrets, encryption key, and SSH host-key seed.
  • Check target trust before reopening access. Reissue legacy SSH and NETCONF checkouts with a verified host key. Check TLS certificates for database targets and enable Network Level Authentication for RDP.
  • If your cluster token was generated from fewer than 32 random bytes, replace it on every role. Move monitoring that uses cluster credentials to a read-only service token.
  • If automation uses a temporary admin checkout to create other checkouts, switch it to an authorized broker or permanent administrator. Give each reviewer an admin or auditor checkout.
  • For Helm chart 3.0.0, use Kubernetes 1.33 or later. The proxy service defaults to a private address on web port 8080. To expose it publicly, configure external TLS and acknowledge public access in the chart. For external PostgreSQL, verify that Bridge trusts the certificate and that it matches the database hostname.
  • In online deployments with a broker token, v3.0.0 enables platform telemetry by default. If your tenant does not support telemetry, disable it before starting Bridge.

New

  • Choose native access, browser access, or both for each checkout.
  • A single Bridge deployment can authenticate users across multiple LDAP directories or Active Directory domains. Configure TLS for each directory when adopting this option. Existing single-directory configurations continue to work.
  • Try VNC desktops and Oracle Database through a browser or compatible native client. Both betas are off by default and are for development and testing only.
  • Use WebSockets through an HTTP proxy checkout. The recording shows the handshake and how much data traveled in each direction. Message contents are excluded.
  • Administrators can recover playable portions of incomplete RDP recordings while preserving the original audit data.
  • With telemetry enabled, reviewers can open the session from its end event in Britive. Review requires an authorized checkout.

Improved

  • Administrators and auditors can watch live sessions across Bridge instances and use the controls their checkout permits.
  • Session audits include PostgreSQL bulk transfers and SQL Server remote procedure calls. CSV exports use spreadsheet-safe text by default, with a raw format available for the original values.
  • HTTP proxy connections close when their checkout expires or is revoked. Reconnecting through a still-active checkout starts a separate recording.
  • Browser database sessions and native Oracle sessions close when their checkout credentials or policy change.
  • Helm enables a separate worker for GIF and video exports by default. If you only use browser playback, you can turn it off.

Fixed

  • HTTP proxy recordings now survive a proxy restart.
  • Closing the browser while a desktop session disconnects no longer leaves its recording marked incomplete.

Links

Last updated on