DBA

Database administration endpoints for on-demand pruning and migration deployment. They complement the automated maintenance schedule: use them for manual cleanup or from CI/CD.

All endpoints require admin access (admin or superadmin role).

Prune on demand

POST /api/dba/prune

Runs HotMesh’s built-in prune operations against the workflow database for one namespace. Every field except app_id is optional, and the defaults are conservative.

Auth: Requires admin access.

Request body

Field	Type	Default	Description
`app_id`	`string`		Required. Application namespace to prune.
`expire`	`string`	`"7 days"`	Retention period. Only records older than this are affected.
`jobs`	`boolean`	`true`	Hard-delete expired jobs.
`streams`	`boolean`	`true`	Hard-delete expired streams.
`engineStreams`	`boolean`		Prune engine consumer streams.
`engineStreamsExpire`	`string`		Retention for engine stream entries.
`workerStreams`	`boolean`		Prune worker consumer streams.
`workerStreamsExpire`	`string`		Retention for worker stream entries.
`attributes`	`boolean`	`false`	Strip execution artifacts (activity inputs and outputs) from jobs.
`entities`	`string[]`		Entity allowlist. Only jobs matching these entity names are pruned.
`pruneTransient`	`boolean`	`false`	Delete transient jobs (where entity is null).
`keepHmark`	`boolean`	`false`	Preserve the `hmark` field during attribute stripping.

Example request (delete everything older than 30 days)

curl -s https://lt.acme.dev/api/dba/prune \
  -X POST \
  -H "Authorization: Bearer $LT_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"app_id": "longtail", "expire": "30 days", "jobs": true, "streams": true, "pruneTransient": true}'

Example request (strip artifacts from review workflows, keep hmark)

curl -s https://lt.acme.dev/api/dba/prune \
  -X POST \
  -H "Authorization: Bearer $LT_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"app_id": "longtail", "expire": "14 days", "attributes": true, "entities": ["reviewContent"], "keepHmark": true}'

Response 200

{
  "jobs": 42,
  "streams": 156,
  "attributes": 0,
  "transient": 18,
  "marked": 0
}

Each field is the count of records that operation affected.

Response 400: { "error": "app_id is required" }

Deploy

POST /api/dba/deploy

Deploys the server-side prune function and runs all pending database migrations for a namespace. Both steps are idempotent, so it is safe to call on every startup or from a CI/CD pipeline.

Auth: Requires admin access.

Request body

Field	Type	Required	Description
`app_id`	`string`	Yes	Application namespace to deploy into.

The call runs dbaService.deploy(), which:

Creates or replaces the HotMesh server-side prune function in Postgres.
Runs migrate() to apply any pending schema files from services/db/schemas/.

Example request

curl -s https://lt.acme.dev/api/dba/deploy \
  -X POST \
  -H "Authorization: Bearer $LT_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"app_id": "longtail"}'

Response 200

{ "ok": true }

Response 400: { "error": "app_id is required" }