Edit retention and settings

curl -X PUT http://localhost:5080/api/v1/streams/$STREAM_ID/settings \
  -H "authorization: Bearer $MS_JWT" \
  -H 'content-type: application/json' \
  -d '{
    "retention_days": 30,
    "settings": { "description": "Application logs", "queryable": true }
  }'

curl -X PUT http://localhost:5080/api/v1/streams/$STREAM_ID/settings \
  -H "authorization: Bearer $MS_JWT" \
  -H 'content-type: application/json' \
  -d '{
    "retention_days": 30,
    "settings": { "description": "Application logs", "queryable": true }
  }'

PUT /api/v1/streams/{id}/settings

Updates the retention and the settings of a stream. Send only the properties to change. The call requires streams.configure. On the _molesignal system streams it requires sys.telemetry.manage and changes only the retention, field indexes, extraction, masking, and downsampling. Body

Field	Type	Notes
`retention_days`	integer or `null`	1 through 3650 days. `null` clears the explicit value, and the stream inherits the platform default from Settings → Data retention. A value outside the range answers `400`.
`settings`	object	Stream settings, see below.
`fields`	array	Field indexes, see Configure field indexes.

Settings:

Setting	Notes
`description`	Free text shown in the web UI.
`queryable`	`false` keeps the stream receiving and retaining data, hides it from query selectors, and makes queries fail with `403` (`stream is not queryable`).
`max_query_range_hours`	The most hours one SQL or PromQL query of the stream may cover, from 1 to 87600. A query over a longer range answers `400`. `null` removes the limit.
`field_masking`	Stream-level overrides of the global field masking rules.
`daily_merge`	Logs and traces only. `true`, the default, merges the files of a finished day into one an hour after the day ends, when together they are no larger than `[compactor].target_size`. Retention then applies by the day. `false` keeps the hourly files.
`downsample`	Metrics only. `{ "after_days": 7, "interval_secs": 300 }` rolls samples older than `after_days` days up into buckets of `interval_secs` seconds: the samples of a series in one bucket become one point, values averaged and delta counts summed, and the original samples are deleted. `after_days` is 0 through 3650, `0` turns downsampling off (the default); `interval_secs` is at least 60 and divides a day evenly.

curl -X PUT http://localhost:5080/api/v1/streams/$STREAM_ID/settings \
  -H "authorization: Bearer $MS_JWT" \
  -H 'content-type: application/json' \
  -d '{
    "retention_days": 30,
    "settings": { "description": "Application logs", "queryable": true }
  }'

The response is the updated stream. Retention and downsampling control when the compactor removes or rolls up data. A change affects later clean-ups and never restores deleted files or samples. The platform default for streams without a retention of their own is read with GET /api/v1/settings/stream_retention and set with PUT and the same body, { "retention_days": 30 }, both with sys.settings.manage in the system scope. retention_days is a whole number from 1 to 3650.

Last modified on October 26, 2026

Overview

Authentication

Intake

Query

APM

RUM

Streams

Alerting

Correlation & Web

Platform & Administration

Edit retention and settings