Update eniz1806/vaults3 Docker tag to v4.4.47 #331

2026-08-01T22:02:08+02:00

renovate-bot commented

2026-08-01 22:02:08 +02:00

This PR contains the following updates:

Package	Update	Change
eniz1806/vaults3	patch	`4.4.45` → `4.4.47`

⚠️ Warning

Some dependencies could not be looked up. Check the Dependency Dashboard for more information.

Release Notes

Kodiqa-Solutions/VaultS3 (eniz1806/vaults3)

`v4.4.47`

Compare Source

Added

VaultS3 now measures its own on-disk footprint, so "how much are my objects
really costing" is answered with a number instead of an inference (issue #43,
follow-up reported by kesavkolla). The dashboard Stats panel, vaults3-cli info
and GET /api/v1/cluster/info present three sizes side by side and say what each
one counts:
- Logical: each object's current version, counted once cluster-wide.
- VaultS3 on disk: what its data, metadata, erasure, cold-tier and Raft
  directories actually occupy, summed per node. This is the figure to compare
  against logical for a real amplification ratio (replicas, parity shards,
  non-current versions), and it is shown as an explicit multiple.
- Filesystems: statfs of the whole volumes, which also counts the operating
  system, container images, logs, and any other tenant of the same disk.
The panel stacks VaultS3's share and everything else on one bar, so the gap
between an object total and a much larger "used" figure is visible rather than
suspicious. A per-directory breakdown separates object data from metadata and
Raft logs, which is what distinguishes real growth from a busy volume.
Sizes are allocated blocks, matching du, so a bucket full of small objects
shows the block-rounding cost it genuinely pays. Hardlinks are counted once and
symlinks are not followed.
The walk runs in the background and is cached: storage.usage_scan_interval_secs
(env VAULTS3_USAGE_SCAN_INTERVAL_SECS, Helm usageScanIntervalSecs, default
300) caps how often it may repeat, 0 disables it. It never runs on a request.
The dashboard and CLI show how old the measurement is.
New Prometheus series: vaults3_disk_usage_bytes{dir=...},
vaults3_disk_usage_files{dir=...}, vaults3_disk_usage_bytes_total and
vaults3_disk_usage_scanned_timestamp_seconds, so physical growth can be graphed
next to the logical vaults3_storage_size_bytes_total.

Fixed

A clustered node addressed every peer as itself when peers shared a host on
different ports, because the membership sync rebuilt each peer's API address
from its Raft address plus this node's API port, discarding the configured
cluster.peer_apis. One node per IP (the usual Kubernetes shape) was unaffected,
but a single-host or host-networked cluster silently sent all inter-node traffic
back to itself: object replicas were never placed on peers despite
replica_count: 3, rebalance moved nothing, and the capacity rollup reported
one node's disk usage once per node. Explicitly configured peer addresses now win
over derived ones; this node's own entry stays derived, since the configured form
is a bind address that may be a wildcard. Existing objects written before the fix
stay single-copy until vaults3-cli cluster rebalance.
The Raft directory was missing from the reported storage directories, so its log
and snapshot growth was invisible in capacity numbers on a clustered node.

Changed

The Stats card previously labelled "Total Storage" is now "Logical Storage", and
the capacity panel names each figure, rather than leaving two very different
numbers labelled "storage" next to each other.
The Stats page had a handful of strings the i18n sweep missed (the auto-refresh
toggle, the capacity legend, per-bucket quota lines); they are translated now.
Settings shows the footprint scan interval.

`v4.4.46`

Compare Source

Added

The dashboard is now translated (issue #33, requested by autool). It ships
English, German, French, and Simplified Chinese, picks a language from the
browser on first load, and offers a switcher in the top bar that remembers the
choice. <html lang> follows the selection, so screen readers pronounce the UI
correctly and browsers stop offering to translate an already-translated page.
- All 517 user-facing strings across the 19 pages and 16 components go through
  a translation lookup, including button labels, table headers, placeholders,
  tooltips, empty states, toasts, and error messages.
- Implemented without adding a dependency: a ~120-line provider next to the
  existing theme provider, plus one flat JSON file per language. The dashboard
  still has three runtime dependencies.
- Adding a language is one JSON file and no code, see
  Translating the dashboard. A key
  a translation is missing falls back to English at runtime, so a partial
  translation never breaks the UI, and npx vitest run checks that every
  shipped locale has all the English keys, none that English does not define,
  and the same {placeholders} in every string.
- The German, French, and Chinese files were drafted without a native-speaker
  review and corrections are welcome as PRs. Server-side messages (S3 API
  errors, log lines) remain English.
- All locales are bundled up front rather than fetched on demand, which costs
  about 17 kB gzipped for the three added languages (123 kB to 140 kB) and
  avoids an untranslated flash on load. Worth revisiting if many more languages
  arrive.
- Helm chart 0.1.6, appVersion 4.4.46.

Configuration

📅 Schedule: (UTC)

Branch creation
- At any time (no schedule defined)
Automerge
- At any time (no schedule defined)

🚦 Automerge: Enabled.

♻ Rebasing: Whenever PR is behind base branch, or you tick the rebase/retry checkbox.

🔕 Ignore: Close this PR and you won't be reminded about this update again.

If you want to rebase/retry this PR, check this box

This PR has been generated by Mend Renovate CLI.

This PR contains the following updates: | Package | Update | Change | |---|---|---| | [eniz1806/vaults3](https://github.com/Kodiqa-Solutions/VaultS3) | patch | `4.4.45` → `4.4.47` | --- > ⚠️ **Warning** > > Some dependencies could not be looked up. Check the [Dependency Dashboard](issues/3) for more information. --- ### Release Notes <details> <summary>Kodiqa-Solutions/VaultS3 (eniz1806/vaults3)</summary> ### [`v4.4.47`](https://github.com/Kodiqa-Solutions/VaultS3/blob/HEAD/CHANGELOG.md#4447---2026-08-01) [Compare Source](https://github.com/Kodiqa-Solutions/VaultS3/compare/v4.4.46...v4.4.47) ##### Added - **VaultS3 now measures its own on-disk footprint**, so "how much are my objects really costing" is answered with a number instead of an inference (issue [#43](https://github.com/Kodiqa-Solutions/VaultS3/issues/43), follow-up reported by kesavkolla). The dashboard Stats panel, `vaults3-cli info` and `GET /api/v1/cluster/info` present three sizes side by side and say what each one counts: - **Logical**: each object's current version, counted once cluster-wide. - **VaultS3 on disk**: what its data, metadata, erasure, cold-tier and Raft directories actually occupy, summed per node. This is the figure to compare against logical for a real amplification ratio (replicas, parity shards, non-current versions), and it is shown as an explicit multiple. - **Filesystems**: `statfs` of the whole volumes, which also counts the operating system, container images, logs, and any other tenant of the same disk. The panel stacks VaultS3's share and everything else on one bar, so the gap between an object total and a much larger "used" figure is visible rather than suspicious. A per-directory breakdown separates object data from metadata and Raft logs, which is what distinguishes real growth from a busy volume. - Sizes are allocated blocks, matching `du`, so a bucket full of small objects shows the block-rounding cost it genuinely pays. Hardlinks are counted once and symlinks are not followed. - The walk runs in the background and is cached: `storage.usage_scan_interval_secs` (env `VAULTS3_USAGE_SCAN_INTERVAL_SECS`, Helm `usageScanIntervalSecs`, default 300\) caps how often it may repeat, `0` disables it. It never runs on a request. The dashboard and CLI show how old the measurement is. - New Prometheus series: `vaults3_disk_usage_bytes{dir=...}`, `vaults3_disk_usage_files{dir=...}`, `vaults3_disk_usage_bytes_total` and `vaults3_disk_usage_scanned_timestamp_seconds`, so physical growth can be graphed next to the logical `vaults3_storage_size_bytes_total`. ##### Fixed - **A clustered node addressed every peer as itself when peers shared a host on different ports**, because the membership sync rebuilt each peer's API address from its Raft address plus *this* node's API port, discarding the configured `cluster.peer_apis`. One node per IP (the usual Kubernetes shape) was unaffected, but a single-host or host-networked cluster silently sent all inter-node traffic back to itself: **object replicas were never placed on peers despite `replica_count: 3`**, rebalance moved nothing, and the capacity rollup reported one node's disk usage once per node. Explicitly configured peer addresses now win over derived ones; this node's own entry stays derived, since the configured form is a bind address that may be a wildcard. Existing objects written before the fix stay single-copy until `vaults3-cli cluster rebalance`. - The Raft directory was missing from the reported storage directories, so its log and snapshot growth was invisible in capacity numbers on a clustered node. ##### Changed - The Stats card previously labelled "Total Storage" is now "Logical Storage", and the capacity panel names each figure, rather than leaving two very different numbers labelled "storage" next to each other. - The Stats page had a handful of strings the i18n sweep missed (the auto-refresh toggle, the capacity legend, per-bucket quota lines); they are translated now. - Settings shows the footprint scan interval. ### [`v4.4.46`](https://github.com/Kodiqa-Solutions/VaultS3/blob/HEAD/CHANGELOG.md#4446---2026-08-01) [Compare Source](https://github.com/Kodiqa-Solutions/VaultS3/compare/v4.4.45...v4.4.46) ##### Added - **The dashboard is now translated** (issue [#33](https://github.com/Kodiqa-Solutions/VaultS3/issues/33), requested by autool). It ships **English, German, French, and Simplified Chinese**, picks a language from the browser on first load, and offers a switcher in the top bar that remembers the choice. `<html lang>` follows the selection, so screen readers pronounce the UI correctly and browsers stop offering to translate an already-translated page. - All 517 user-facing strings across the 19 pages and 16 components go through a translation lookup, including button labels, table headers, placeholders, tooltips, empty states, toasts, and error messages. - Implemented without adding a dependency: a \~120-line provider next to the existing theme provider, plus one flat JSON file per language. The dashboard still has three runtime dependencies. - **Adding a language is one JSON file and no code**, see [Translating the dashboard](CONTRIBUTING.md#translating-the-dashboard). A key a translation is missing falls back to English at runtime, so a partial translation never breaks the UI, and `npx vitest run` checks that every shipped locale has all the English keys, none that English does not define, and the same `{placeholders}` in every string. - The German, French, and Chinese files were drafted without a native-speaker review and corrections are welcome as PRs. Server-side messages (S3 API errors, log lines) remain English. - All locales are bundled up front rather than fetched on demand, which costs about 17 kB gzipped for the three added languages (123 kB to 140 kB) and avoids an untranslated flash on load. Worth revisiting if many more languages arrive. - Helm chart 0.1.6, appVersion 4.4.46. </details> --- ### Configuration 📅 **Schedule**: (UTC) - Branch creation - At any time (no schedule defined) - Automerge - At any time (no schedule defined) 🚦 **Automerge**: Enabled. ♻ **Rebasing**: Whenever PR is behind base branch, or you tick the rebase/retry checkbox. 🔕 **Ignore**: Close this PR and you won't be reminded about this update again. --- - [ ] If you want to rebase/retry this PR, check this box --- This PR has been generated by [Mend Renovate CLI](https://github.com/renovatebot/renovate).

renovate-bot added 1 commit 2026-08-01 22:02:08 +02:00

Update eniz1806/vaults3 Docker tag to v4.4.47 754c05f8b3

renovate-bot scheduled this pull request to auto merge when all checks succeed 2026-08-01 22:02:09 +02:00

renovate-bot merged commit a1a23839b7 into main

2026-08-01 22:02:10 +02:00

renovate-bot deleted branch renovate/eniz1806-vaults3-4.x

2026-08-01 22:02:10 +02:00

Sign in to join this conversation.

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: arnoutvw/compose-files#331