Connections: deleting a core does not offer to delete its local data, and logs/backups/json references keep it after a Storage delete

[Alena-Selezneva]

Status (after #748)

Deleting a core's local data per core now exists in Settings → General → Storage → "Data by core" (reports, strategy history, order traces, warning episodes). What remains for this issue:

• offer the delete in the Connections delete flow itself;
• logs/, backups/settings/, backups/strategies/, panic.log, and uid references in figures.json, layout.toml, charts.json, core_groups, Telegram chat_access, tab_badges.json, core_updates.json are not touched by any delete;
• server-wide warning episodes (keyed by server IP) stay.

The original report follows.

---

User request

@A1S2M3S4Q9 in the MoonTerminal chat (2026-09-25): «как правильно удалить на локальной машине уже подключенные ядра, чтобы на локальной машине не осталось и следа, что было что-то подключено?» (how do I remove already-connected cores so that no trace is left on the local machine).

What happens today (checked in code)

Settings → Connections → «Удал» → «Сохранить» removes the core from servers.enc, stops its session and prunes its chart specs (settings/connections/table.rs::delete_server, settings/apply.rs). Nothing else is touched. What stays on disk for that core:

• its rows in the report replica reports.sqlite (by design its core_uid is never reused, see db::max_core_uid — so the rows are never overwritten either);
• strategies.sqlite (strategy version history), order_traces.sqlite, core_warnings.sqlite — kept deliberately as the only copy;
• daily snapshots under backups/settings/ still contain the old servers.enc with the removed core (encrypted);
• logs/ (core addresses appear in log lines);
• panic.log in the working directory.

The only way to wipe it today is to delete the whole data directory plus the moon-terminal OS keyring entry, which also wipes every other core.

Proposal

When deleting a core, offer «also delete this core's local data» (report rows, strategy history, order traces, warnings for this core_uid), ideally with a preview of what will be removed, like the storage-cleanup preview in «Настройки → Хранилище». Backups/logs could be mentioned in the dialog as not covered, or pruned too.

Use cases: selling/handing over a VPS or a core, a shared machine, or just not wanting history of a core you no longer use.

[Alena-Selezneva]

Related: #553 covers one slice of this. A deleted core's report rows keep it in Analytics and the Profit Monitor. The "forget this core's data" option asked for here would resolve #553 as well, so the two are worth doing together.

[guyverino]

Partly done in #748 (merged as c9e0a53).

Landed: Settings → General → Storage → "Data by core" lists every core the local databases hold data for — deleted connections included — with per-core counts, and deletes a core's local data on request: report rows (reports.sqlite, with valuations, clock-offset segments and sync metadata for a core no longer in Connections), strategy history (strategies.sqlite), order traces (order_traces.sqlite) and its own warning episodes (core_warnings.sqlite). The confirmation names the cores and the counts.

Still open (this issue now tracks only this):

• the option is not offered in the Connections delete flow itself — a user deleting a core there still has to go to Storage;
• not covered by any delete: logs/ (core addresses in log lines), backups/settings/ and backups/strategies/ snapshots, panic.log, and uid references in figures.json, layout.toml (recent_coins, header_ticker), charts.json, core_groups, Telegram chat_access, tab_badges.json, core_updates.json;
• server-wide warning episodes (keyed by server IP, not by core) stay, as they cannot be attributed to one core on a shared server.