VersusControl/versus-incident v1.4.34

on GitHub

Linked Logs for Kubernetes Breaks

• Combine bounded crash excerpts with logs from the linked data source in one
  investigation view, ordered by time and without duplicate lines.
• Open a break before all logs have arrived. See pending coverage and let the
  view retry automatically as the source catches up, without moving the
  selected break or its time window.
• Use source-specific ingestion delays by default, with an optional override
  in Kubernetes agent settings.
• Keep bounded settled excerpts available after unlinking the source and
  restarting, subject to current cluster permissions.
• Preserve the correct Pod, container run and OOMKilled cause as workloads
  restart. Show missing or budget-limited evidence instead of false completeness.
• Improve break evidence without storing a continuous container-log stream.

Stable Kubernetes Break Investigations

• Open the break you selected, including recurring and retained occurrences,
  without silently switching to another break for the same resource.
• Keep the log window, bracket, selected marker and filters stable as evidence
  loads.
• Keep Logs at break visible when timestamps or captures are unavailable, with
  clear explanations instead of a disappearing panel.
• Read focused changes and crashes without detail-level paging. Incomplete
  coverage remains explicit; cluster Timeline and other lists still paginate.

Workloads That Stay Useful During Sync

• Keep authorized inventory and known counts visible during temporary refresh
  failures, including the first usable partial inventory.
• Show unknown counts as unavailable, not as misleading zeros. Successfully
  observed empty inventory still shows zero.
• Recover automatically from temporary preparation delays, with an explicit
  retry when automatic recovery is exhausted.
• Remove deleted resources after a committed refresh and prevent old scoped
  snapshots from hiding newer collection failures.
• Offer to restart paging when inventory changes invalidate a page, rather than
  repeating previously shown rows.

Clearer Workload Layout

• Align headers and row values across mixed resource kinds, including rows with
  and without an immediate Logs action.
• Preserve compact mobile layouts, readable names and separate selection and
  Logs controls.

Kubernetes History Ready on Upgrade

• Open Timeline, Changes, Crashes and historical Issues quickly as soon as
  Versus starts after an upgrade. Saved history is prepared at startup, with
  progress in the log.
• Keep very large histories preparing in the background and show a clear,
  retryable "history upgrading" status instead of a timeout.
• Start several replicas at once during an upgrade without them stalling each
  other.
• Read clearer diagnostics that separate storage problems from data problems.
• Follow a Postgres storage guide that explains what each history setting
  controls and how it affects the database.

Enterprise

• Share the same investigation and Workloads improvements while preserving
  existing cluster and namespace permissions.
• Keep the one-organization self-hosted upgrade model: licensed writes and
  ordered reads of licensed and retained OSS evidence.
• Continue clearing protected data when access is revoked; stored evidence does
  not bypass current authorization or physical-cluster identity checks.
• Prepare licensed and retained OSS history at startup, under the same
  organization and permission rules.

Scope

• Adds an optional linked-log ingestion-delay setting; existing data-source
  connections are reused.
• Retires the kubernetes.evidence.read_models setting; an existing value is
  ignored. The first start after upgrading may take longer while history is
  prepared. No saved history is deleted.
• Does not add continuous log storage, a learning worker or a general history
  retention policy change.
• Old capture records stranded under pre-fix recovery scopes are not
  automatically migrated.
• Licensed Enterprise runtime acceptance remains unverified.
• This is a changelog draft, not a published release announcement.

What's Changed

• fix(kubernetes): keep investigations and workloads stable during sync by @hoalongnatsu in #396
• feat(kubernetes): enrich break investigations with linked logs by @hoalongnatsu in #397
• feat(kubernetes): keep investigation history fast and available across upgrades by @hoalongnatsu in #398
• docs(aws): add docs for aws harness dashboard by @hoalongnatsu in #399

Full Changelog: v1.4.33...v1.4.34