feat(audit-log): show resolved names instead of UUIDs in the table + detail

YousifShkara commented

2026-06-10 05:52:20 +02:00

Owner

The audit log table rendered an 8-char UUID prefix in both the "Entity ID" and "User" columns, and the expanded detail row dumped the full UUIDs again. Every other user-visible reference to an actor or a record in mokosh already uses resolved names (assigned_to_name on tickets, display_name() on calendar, user_name() on projects); the audit log was the last surface where end users saw a raw UUID.

Pairs with the mokosh-server patch (feat/audit-log-resolve-names) that adds user_name: Option<String> and entity_name: Option<String> to AuditLogEntryResponse, computed server-side via a LEFT JOIN against users plus a CASE-per-entity-type subselect (companies/contacts/sites/contracts/contract_items/rate_cards/invoices/tickets/projects/assets/credential_vault/payment_gateway_configs/tax_rates, plus the auth synthetic type whose entity_id is itself a user_id).

SPA wiring:

AuditLogEntry gains the two new Option fields with #[serde(default)] so an older server build (pre-rename, before the JOIN ships) decodes cleanly and the UI falls back to the short-UUID rendering rather than failing to parse the response.
The table row reads entity_name and user_name. user_name falls back to "System" when null (catches system-issued rows + JIT-created users whose name hasn't been onboarded yet); entity_name falls back to the short-UUID prefix so a brand-new entity_type that doesn't have a server-side resolver row yet still gets SOMETHING the reader can correlate against.
The "Entity ID" column header is renamed "Record" - the column now shows a human label, not an identifier, and the "Entity" column to its left already carries the type label.
The expanded detail view renders the resolved name as the primary value, with the raw UUID demoted to a small muted sub-line for traceability when correlating across logs/DB. The sub-line is suppressed entirely when the underlying UUID is null (system-issued rows).

Net effect for the user: the audit log table now reads "yousif (Yousif Lastname) created ticket T000042 in tenant X" instead of "8abc12... created entity in tenant 4f3e2d...". The short UUID remains as a safety net for any entity_type the server doesn't yet resolve, and a small muted UUID sub-line in the detail view preserves the ability to chase a row through logs by id when debugging.

The audit log table rendered an 8-char UUID prefix in both the "Entity ID" and "User" columns, and the expanded detail row dumped the full UUIDs again. Every other user-visible reference to an actor or a record in mokosh already uses resolved names (assigned_to_name on tickets, display_name() on calendar, user_name() on projects); the audit log was the last surface where end users saw a raw UUID. Pairs with the mokosh-server patch (feat/audit-log-resolve-names) that adds `user_name: Option<String>` and `entity_name: Option<String>` to `AuditLogEntryResponse`, computed server-side via a LEFT JOIN against users plus a CASE-per-entity-type subselect (companies/contacts/sites/contracts/contract_items/rate_cards/invoices/tickets/projects/assets/credential_vault/payment_gateway_configs/tax_rates, plus the `auth` synthetic type whose entity_id is itself a user_id). SPA wiring: - `AuditLogEntry` gains the two new Option<String> fields with `#[serde(default)]` so an older server build (pre-rename, before the JOIN ships) decodes cleanly and the UI falls back to the short-UUID rendering rather than failing to parse the response. - The table row reads `entity_name` and `user_name`. `user_name` falls back to "System" when null (catches system-issued rows + JIT-created users whose name hasn't been onboarded yet); `entity_name` falls back to the short-UUID prefix so a brand-new entity_type that doesn't have a server-side resolver row yet still gets SOMETHING the reader can correlate against. - The "Entity ID" column header is renamed "Record" - the column now shows a human label, not an identifier, and the "Entity" column to its left already carries the type label. - The expanded detail view renders the resolved name as the primary value, with the raw UUID demoted to a small muted sub-line for traceability when correlating across logs/DB. The sub-line is suppressed entirely when the underlying UUID is null (system-issued rows). Net effect for the user: the audit log table now reads "yousif (Yousif Lastname) created ticket T000042 in tenant X" instead of "8abc12... created entity in tenant 4f3e2d...". The short UUID remains as a safety net for any entity_type the server doesn't yet resolve, and a small muted UUID sub-line in the detail view preserves the ability to chase a row through logs by id when debugging.

YousifShkara self-assigned this

2026-06-10 05:52:20 +02:00

YousifShkara added 1 commit

2026-06-10 05:52:21 +02:00

feat(audit-log): show resolved names instead of UUIDs in the table + detail

Create release / Create release from merged PR (pull_request) Has been skipped

Details

Check / clippy + fmt + tests (pull_request) Successful in 1m59s

Details

841e47ad48

The audit log table rendered an 8-char UUID prefix in both the "Entity ID" and "User" columns, and the expanded detail row dumped the full UUIDs again. Every other user-visible reference to an actor or a record in mokosh already uses resolved names (assigned_to_name on tickets, display_name() on calendar, user_name() on projects); the audit log was the last surface where end users saw a raw UUID.

Pairs with the mokosh-server patch (feat/audit-log-resolve-names) that adds `user_name: Option<String>` and `entity_name: Option<String>` to `AuditLogEntryResponse`, computed server-side via a LEFT JOIN against users plus a CASE-per-entity-type subselect (companies/contacts/sites/contracts/contract_items/rate_cards/invoices/tickets/projects/assets/credential_vault/payment_gateway_configs/tax_rates, plus the `auth` synthetic type whose entity_id is itself a user_id).

SPA wiring:

- `AuditLogEntry` gains the two new Option<String> fields with `#[serde(default)]` so an older server build (pre-rename, before the JOIN ships) decodes cleanly and the UI falls back to the short-UUID rendering rather than failing to parse the response.
- The table row reads `entity_name` and `user_name`. `user_name` falls back to "System" when null (catches system-issued rows + JIT-created users whose name hasn't been onboarded yet); `entity_name` falls back to the short-UUID prefix so a brand-new entity_type that doesn't have a server-side resolver row yet still gets SOMETHING the reader can correlate against.
- The "Entity ID" column header is renamed "Record" - the column now shows a human label, not an identifier, and the "Entity" column to its left already carries the type label.
- The expanded detail view renders the resolved name as the primary value, with the raw UUID demoted to a small muted sub-line for traceability when correlating across logs/DB. The sub-line is suppressed entirely when the underlying UUID is null (system-issued rows).

Net effect for the user: the audit log table now reads "yousif (Yousif Lastname) created ticket T000042 in tenant X" instead of "8abc12... created entity in tenant 4f3e2d...". The short UUID remains as a safety net for any entity_type the server doesn't yet resolve, and a small muted UUID sub-line in the detail view preserves the ability to chase a row through logs by id when debugging.

YousifShkara merged commit bfd0b317ed into main

2026-06-10 05:52:24 +02:00

YousifShkara deleted branch feat/audit-log-show-names

2026-06-10 05:52:24 +02:00

YousifShkara referenced this pull request from a commit

2026-06-10 05:52:25 +02:00

Merge pull request 'feat(audit-log): show resolved names instead of UUIDs in the table + detail' (#107) from feat/audit-log-show-names into main

Rows
Columns

feat(audit-log): show resolved names instead of UUIDs in the table + detail #107