Arius Web UI — user guide¶

The Arius Web UI is a single-page app for archiving and restoring files to Azure Blob Storage through your browser — no command line needed. It connects to a storage account, reads the encrypted manifest already in your container, and lets you browse snapshots, run archive and restore jobs, watch them live, and search across every repository you've added.

This guide is task-first: what you click and what happens. For how it's built (the Arius.Api + SignalR backend behind the SPA), see the Web host design doc. For how to run the server in Docker or in dev, see the deployment guide and the repository README. Unfamiliar terms (snapshot, chunk, tier, pointer, rehydration) are defined in the glossary.

Heads-up about the passphrase. Every chunk is encrypted with your repository passphrase. Arius cannot recover it. If you lose it, the data in the container is unrecoverable. Store it somewhere safe before you archive anything.

The layout¶

A narrow icon rail on the left switches between the four top-level areas:

Rail item	What it is
Overview	KPI cards, a Storage accounts table, and the Repositories table — whose header carries the Add existing / New repository buttons. Your home screen.
Repos	The repositories list — same data.
Jobs	Every archive/restore run (one-off and scheduled) with a live output console.
Settings	Placeholder for now.

The top bar shows a breadcrumb and — on every screen except Overview — a global search box. Press ⌘K (macOS) or Ctrl+K (Windows/Linux) anywhere to open search; Esc closes it.

First run — add a repository¶

Open the app, go to Overview (or Repos), and choose one of the two buttons in the Repositories card header. Both are two-step wizards; the difference is whether the container already exists.

Add existing — point at a container that already holds an Arius archive. Arius reads the existing manifest and snapshots; no files are uploaded.
New repository — create a brand-new container to archive into.

Step 1 — the storage account (both wizards)¶

Pick Use configured to reuse a storage account you've already added (each row shows how many repositories use it), or Add new account and enter:

Account name — your Azure Storage account name.
Account key — the account key. It's stored encrypted server-side (Data-Protection) and is never sent back to the browser.

Mandatory fields are marked with a red asterisk. (There is no region to pick here — the region used for cost estimates is read from the container's own metadata; see Managing storage accounts.)

In the Add existing wizard, the button is Connect & discover: Arius connects with that account and live-streams the container names it finds. The New repository wizard uses Continue (it doesn't need to list containers — it's making a new one).

Step 2a — Add existing: pick the container¶

You'll see the discovered containers as a radio list. Select the one holding your archive, then fill in:

Friendly alias (optional) — the display name used throughout the UI; leave it blank to default to the container name.
Passphrase — the repository passphrase (needed to read/write encrypted chunks).
Local path (optional) — a folder on the server that the Files view overlays against the archive (and the default source folder for archive runs). It must resolve on the machine running Arius.Api — in Docker that means a folder you've mounted into the container. Click the […] button to browse the folders the server can see, or type the path. See deployment for how the mounts work.

Click Add repository — you land on the repository's Files view.

Step 2b — New repository: create the container¶

Container name (required) — the blob container to create, typed by you (e.g. arius-photos). One container per repository. It's the first field you fill.
Friendly alias (optional) — display name; defaults to the container name when left blank.
Local path (optional) — same meaning as above, with the same […] server-folder browser.
Default tier — Hot, Cool, Cold, or Archive. This is the tier preselected for archive runs; you can still override per run. Archive is the cheapest to store and the slowest/most expensive to read back (it requires rehydration — see below).
Passphrase + Confirm passphrase — must match. The form won't enable Create repository until the container name, passphrase, and confirmation are all filled and matching. Mandatory fields are marked with a red asterisk, and a warning reminds you the passphrase cannot be recovered.

Click Create repository to land on its Files view.

Managing storage accounts¶

The Overview lists every storage account under management in its own table — its name and how many repositories use it. Click an account row to open its edit flyout (a right-hand drawer):

Account key — paste a new key to rotate it (stored encrypted; leave blank to keep the current one). The key belongs to the account, so this is the one place to change it — it is no longer on a repository's Properties.
Delete account (danger zone) — removes an account that has no repositories.

New accounts are added from the Add new account step of either wizard (above).

Where does the pricing region come from? Cost estimates are priced for an Azure region, but Arius reads that region from each container's own metadata — it is not chosen in the UI. Set the container's region metadata (e.g. westeurope) in Azure Storage Explorer for an accurate estimate; if it is unset, Arius prices against a default region (northeurope). The region is used only to pick rates; the resolved region is shown (read-only) in the repository list, flagged (default) when the container's metadata is unset.

Browsing a repository¶

Selecting a repository (from a table row, or after a wizard) opens the repository detail screen: a header with the alias, the container and local-path chips, and the Restore, Archive, and Properties buttons, then the snapshot bar, then two tabs — Files and Statistics. (Restore and Archive open right-hand drawers; Properties opens a drawer too — see Properties below.)

Snapshot / time-travel bar. Each archive run produces a snapshot. The bar sits above the tabs and applies to whatever tab you're on. It shows the active snapshot with a LATEST pill, a dropdown picker listing every snapshot newest-first (version + timestamp), and a scrubber with one dot per snapshot. Versions are numbered oldest-first — v1 is the first archive run, the highest number is the latest. Click a dropdown item or a scrubber dot to time-travel — the Files explorer then shows the repository exactly as it was at that snapshot, and an amber Historical view badge appears. Pick the latest snapshot (the top of the dropdown, or LATEST) to return to the live working state.

Files tab — the explorer¶

The Files tab is where you browse what's in the repository and pick files to restore, scoped to the snapshot selected in the bar above.

Explorer (folder tree + file list). The left pane is the folder tree (expand folders to drill in); the right pane lists the files in the selected folder. Above them: an up button, the current path, and a filter box that matches file names in the current folder (case-insensitive substring). Each file row shows:

State ring + label — a small two-sided icon: the left half is your local disk, the right half is the repository. Outer ring = pointer/filetree entry; inner disc = the actual binary/chunk. Colors: dark = present, blue = binary-on-disk / chunk hydrated, light blue = chunk not hydrated (sitting in the archive tier), grey = absent. Click State legend in the footer for the full key. The text label summarises it: In sync, Pointer only, In repository, Not archived, Archive tier, Rehydrating.
Size, Tier (Online or Archive), and Modified date.

Collecting files to restore. Click any file row to collect it (a checkmark fills in, the row highlights). A blue collected bar appears showing the count and total size, with Clear and Restore collected buttons. Collecting respects the snapshot you're viewing — collect from a historical snapshot to restore that older version. Use Restore collected to open the Restore drawer scoped to exactly those files.

Statistics tab¶

Five KPI cards in one row, in two scoped sections. This snapshot (follows the snapshot bar) shows Files and Original size — the logical, uncompressed size of the selected snapshot, what you would restore. Repository storage · across all snapshots shows Deduplicated size (unique data before compression), Stored size (the actual cloud footprint after dedup + compression), and Unique chunks. Below the cards, a Stored size by tier breakdown shows each tier's share as a 100%-stacked bar and, per tier, its stored size, chunk count, and estimated monthly cost, with a grand-total monthly cost — priced for the repository's region (read from the container's metadata; see Managing storage accounts).

The two snapshot cards load immediately; the three repository-storage cards (and the cost breakdown) are calculated across all snapshots and load separately. The result is cached serverside per snapshot.

Properties¶

Properties is a right-hand drawer, opened from the button in the repository header.

Friendly alias and Local folder — editable (the local folder has a […] server-folder browser); Save changes persists them.
Storage account and Container — read-only. The account key belongs to the account, not the repository — rotate it from the account's flyout on the Overview (see Managing storage accounts).
Passphrase — type a new passphrase to rotate it; a Confirm passphrase field appears and must match before Save changes is enabled. Leaving it blank keeps the current one. Note that existing snapshots stay readable with the passphrase they were written with — rotating affects future writes, it does not re-encrypt the archive.
Scheduled archives — add a cron expression (e.g. 0 2 * * *, interpreted in UTC) to fire archive runs automatically. Existing schedules show their next run time and a trash button to delete them.
Delete repository (danger zone) — removes the repository from Arius after a confirm step. The Azure container and its blobs are not deleted; you can re-add it later with Add existing.

Archiving (backing up)¶

Click Archive in the repository header to open the Archive drawer (a right slide-over).

Source folder — shown read-only; it's the repository's local path. If it's empty, set one in Properties first (there's nothing local to archive otherwise).
Upload tier — Hot / Cool / Cold / Archive (defaults to the repository's default tier). This is where the newly uploaded chunks land.
On disk after archive — a three-way choice:
Keep files only — upload, write no pointer sidecars (the default).
Keep files + pointers — also write .pointer.arius sidecars next to the binaries (--write-pointers).
Replace with pointers — after a successful upload, delete the local binaries and leave only the pointer sidecars to reclaim disk space (--remove-local, which always writes pointers).
Fast hash — skip re-reading files the local hashcache verifies as unchanged since the last run; faster on large, stable trees, off by default (--fast-hash).
Click Start archive. The drawer switches to the live stream (see Watching live progress). Small files are bundled automatically (1 MB threshold, ~64 MB tar target).

Restoring (downloading)¶

Open the Restore drawer in one of two ways:

Restore in the repository header → restores the whole repository at the latest snapshot.
Restore collected in the Files tab → restores only the files you collected, at the snapshot you were viewing.

The drawer shows the target (whole repository or N collected files) and the snapshot, then two options:

Overwrite existing files — re-restore files even if they're already present at the destination. Off by default (identical files are skipped).
Skip pointer files — don't write pointers.

Click Start restore.

The restore cost confirmation¶

Restoring can cost money: reading from the Cool or Cold tiers incurs a per-GB data-retrieval charge, Archive chunks must be rehydrated first (this takes hours and is pricier), and downloading data out of Azure incurs egress. Whenever a restore has a non-zero estimated cost, Arius pauses before downloading anything and shows a cost-approval modal. The estimate includes data retrieval, operations, and internet egress (under Azure pricing the first 100 GiB/month of egress is free), priced for the repository's region (read from the container's metadata).

The modal adapts to what's being restored:

Archive restore — shows how many chunks must be rehydrated and the size from archive, and asks you to choose a rehydration priority: Standard (~15 h, cheaper) or High (~1 h, pricier), each with its estimated total.
Online restore (Hot/Cool/Cold only) — no rehydration; shows the chunk count, download size, and a single estimated total.

Click Approve & restore to proceed, or Decline to cancel. Closing the drawer (or losing the connection — e.g. closing the tab) while the modal is up also declines and cancels the restore. A genuinely free restore skips the modal entirely.

Watching live progress¶

Once an archive or restore starts, the drawer streams progress over a live WebSocket connection:

A progress bar with a percentage and a status line (Archiving… / Restoring… / Awaiting cost approval / Done).
A stats grid of live counters (files, bytes, chunks — whatever the run reports).
A live console of log lines, scrolling as the job runs.
On completion, a green check with a one-line summary.

You can Hide the drawer while a job runs (it keeps going) and Close it once it's done.

The Jobs screen¶

The Jobs rail item shows every run across all repositories in one table — repository, kind (archive/restore), trigger (manual or schedule), status, and a progress bar. Status badges: Running, Rehydrating, Scheduled, Queued, Completed, Failed. Counters at the top show how many are running and scheduled. Below the table, a Live output console aggregates log lines from active jobs — handy for keeping an eye on a scheduled archive that fired without you.

Global search¶

Press ⌘K / Ctrl+K (or click the search box in the top bar) to open the search overlay. Type to search file names and paths across every repository you've added — matching is a recursive filename filter run per repository, so one unreachable repository won't break the rest of the results. Each hit shows the file's state ring, name, owning repository, path, and size. Click a result to jump to that repository's Files view. Esc closes the overlay.

Where to go next¶

Hosts overview — how the Web UI compares to the CLI and Explorer hosts, and which to use when.
Deployment guide — run the server (Docker single-container, Compose, dev).
Web host design — the Arius.Api + SignalR architecture behind this UI.
Glossary — snapshot, chunk, dedup, tier, pointer, rehydration.
Repository README — project overview and quick start.