release: Telegraf Controller 1.2.0 release notes

content/telegraf/controller/reference/release-notes.md

@@ -9,10 +9,176 @@
 weight: 101
 ---
 
-## v1.1.2 {date="2026-10-06"}
+## v1.2.0 {date="2026-10-27"}
 
 <!-- Update and move the link to the latest version. -->
-[Download Telegraf Controller v1.1.2](/telegraf/controller/install/#download-and-install-telegraf-controller)
+[Download Telegraf Controller v1.2.0](/telegraf/controller/install/#download-and-install-telegraf-controller)
+
+### Before you upgrade
+
+#### Audit logging changes
+
+> [!Important]
+> #### Audit logging stops until you configure a signing key
+>
+> Starting in v1.2.0, {{% product-name %}} signs each audit entry with a
+> key that you configure, and audit logging stays disabled until a key is
+> set. Installations that run with `AUDIT_ENABLED=true` stop writing audit
+> entries after the upgrade until you set `AUDIT_HMAC_KEY`,
+> `AUDIT_HMAC_KEY_FILE`, or `--audit-hmac-key-file`. Existing audit files
+> remain readable.
+> See [Enable audit logging](/telegraf/controller/admin/audit-logs/enable-configure/#enable-audit-logging).
+
+**Reading the audit log requires the new `audit:read` permission.**
+Administrator API tokens created before v1.2.0 do not include it.
+Create a new token to read the audit log with a token.
+Owner tokens with full access are unaffected.
+
+#### Duplicate configuration and group names are renamed
+
+Configuration and configuration group names are now unique.
+On the first start after the upgrade, {{% product-name %}} renames existing
+duplicates: the oldest keeps its name and later ones get ` (2)`, ` (3)`,
+and so on.
+Agents reference configurations by ID or alias, so renamed configurations
+keep serving the same agents.
+See [Create a configuration](/telegraf/controller/configs/create/#create-a-configuration).
+
+#### API token access changes
+
+**Some API tokens lose access they had before.**
+A request authenticated with an API token now manages only its own user's
+tokens and cannot reassign tokens, and account self-service and ownership
+transfer routes accept only a signed-in session.
+Review scripts that use API tokens for these operations.
+The affected endpoints are listed under [API changes](#api-changes).
+
+#### API changes
+
+- `GET /api/heartbeat/agent/:agentId/latest-heartbeat` and
+  `GET /api/heartbeat/agent/:agentId/logs` moved to
+  `GET /api/agents/:id/heartbeat/latest` and
+  `GET /api/agents/:id/heartbeat/logs`. They identify the agent by its
+  {{% product-name %}} `id` and require `agents:read`.
+  See [Retrieve heartbeat data with the API](/telegraf/controller/agents/heartbeat-data/#retrieve-heartbeat-data-with-the-api).
+- `GET /api/heartbeat/agent-status-check` is now
+  `POST /api/agents/status-check` and requires `agents:write`.
+- `PATCH /api/users/me`, `PATCH /api/users/me/password`, and the
+  `/api/users/transfer-owner` routes return `403` to API tokens.
+  See [Session-only endpoints](/telegraf/controller/reference/authentication-authorization/#session-only-endpoints).
+- Token management over an API token is limited to the token owner's own
+  tokens, and reassignment requires an Owner or Administrator session.
+  See [Token visibility](/telegraf/controller/tokens/#token-visibility).
+- `GET /api/password-resets` no longer returns the reset `code`.
+- Reporting rule thresholds must be whole seconds from `1` to
+  `2147483647`. Other values return `400`.
+
+### Features
+
+- Add [security advisories](/telegraf/controller/security-advisories/)
+  (Telegraf Enterprise): {{% product-name %}} downloads the signed
+  InfluxData advisory feed, or imports it offline, matches advisories to
+  the Telegraf version each agent reports, and shows affected agents on the
+  **Security** page, the **Agents** page, and agent detail pages. It also
+  checks its own version against Controller advisories and notifies you
+  when new advisories arrive.
+- Add [pre-shared tokens](/telegraf/controller/tokens/preshared/): supply
+  your own token value when creating a token so one credential can be
+  installed on several {{% product-name %}} instances. Every token now
+  records its source (system-generated, user-provided, or env/flag-provided).
+  Operators can [disable pre-shared tokens](/telegraf/controller/tokens/preshared/#disable-pre-shared-tokens)
+  at startup.
+- Add [bootstrap owner tokens](/telegraf/controller/tokens/bootstrap/):
+  provision an all-access token at startup with `OWNER_TOKEN` or
+  `--owner-token` for infrastructure-as-code workflows. The token is
+  idempotent across restarts and rotates its predecessor when the value
+  changes.
+- Sign audit entries with a configured key and verify the chain
+  (Telegraf Enterprise): each entry records the fingerprint of the key that
+  signed it, {{% product-name %}} verifies the current month's chain at
+  startup, and `GET /api/audit-logger/verify` checks every retained monthly
+  file on demand.
+  See [Verify integrity](/telegraf/controller/admin/audit-logs/view/#verify-integrity).
+- Redesign the [Agents page](/telegraf/controller/agents/monitor/): a fleet
+  status bar with **Fleet Health**, **Telegraf Versions**, and
+  **Last Reported** summaries, removable filter chips, a **Telegraf Version**
+  column, and compact and detailed row layouts. Search now matches the
+  agent's ID, hostname, IP address, or Telegraf version, with
+  [regular expression and match-case toggles](/telegraf/controller/agents/monitor/#search-the-agent-list).
+- Redesign the agent detail page around
+  [heartbeat data](/telegraf/controller/agents/heartbeat-data/): per-heartbeat
+  statistics charts and the log entries reported with the latest heartbeat,
+  with auto-refresh and relative last-reported times.
+- Add a [heartbeat service health](/telegraf/controller/admin/monitor/#heartbeat-service-health)
+  indicator to the navigation and `GET /api/heartbeat/health`, which report
+  why the heartbeat service is unhealthy, not only whether it responds.
+- Include log entries in the default heartbeat output template.
+  New configurations ship with `"logs"` in the plugin's `include` list and a
+  limit of 50 entries per heartbeat.
+  See [Include logs with heartbeats](/telegraf/controller/agents/heartbeat-data/#include-logs-with-heartbeats).
+- Report an agent's real IP address from behind a proxy with the
+  `Telegraf-Agent-IP` heartbeat header.
+  See [Report the agent IP address from behind a proxy](/telegraf/controller/agents/create/#report-the-agent-ip-address-from-behind-a-proxy).
+- Publish {{% product-name %}} as a Docker image on Docker Hub
+  (`influxdata/telegraf-controller`). The image is built for both x86-64
+  and ARM64, so it runs on Intel, AMD, and ARM hosts, including Apple
+  silicon.
+  See [Run {{% product-name %}} in Docker](/telegraf/controller/install/docker/).
+- Ship a signed SHA-256 digests file with every release so you can verify
+  downloaded binaries.
+- Add plugin support to the Telegraf Builder UI:
+  - Google Cloud Secret Manager (`secretstores.googlecloud`)
+
+### Bug fixes
+
+- Stop SQLite database corruption under concurrent agent heartbeats and API
+  or web interface use. The heartbeat service now shares one SQLite library
+  with the rest of {{% product-name %}} instead of linking its own.
+- Enforce the agent limit across a whole
+  [high-availability cluster](/telegraf/controller/admin/high-availability/)
+  instead of per node, and free an agent's slot as soon as it is deleted
+  instead of at the next restart.
+- Keep status checks working with reporting rules of about 25 days or more
+  on PostgreSQL. Previously, one such rule stopped every status check with an
+  integer overflow, so no agent was marked **Not Reporting** or auto-deleted.
+- Reject reporting rule thresholds of zero, negative, or fractional seconds,
+  and ignore rules already stored with a non-positive threshold, which marked
+  live agents **Not Reporting** or deleted them.
+- Keep exactly one default reporting rule. Turning **Default Rule** off on
+  the default rule now returns an error instead of leaving new agents
+  without a rule, and making another rule the default moves the flag in one
+  step.
+- Enforce unique configuration and configuration group names and keep
+  generated names, such as duplicates, within 255 characters. Previously, an
+  overlong generated name broke the configuration list with a `500`
+  response.
+- Limit API tokens to their own permissions: a wildcard such as `*:*` or
+  `license:*` now expands to the permissions the role holds, tokens cannot
+  manage other users' tokens, and the former owner's tokens are re-clamped
+  after an ownership transfer.
+- Forward sign-in, sign-out, constant, and agent-deletion audit events to
+  syslog, webhook, and file destinations like every other event, write
+  nothing while audit logging is disabled, and keep running when a forwarder
+  fails. Previously, disabling audit logging with a syslog forwarder
+  configured could stop the process.
+- Stop one malformed heartbeat from dropping the rest of its batch: reject
+  control characters in `hostname` and `version` with `400`, accept an agent
+  that references the same configuration by ID and by alias, and skip an
+  oversized log entry instead of discarding all entries.
+- Write every acknowledged heartbeat to the database before shutting down.
+- Reject sign-in `returnTo` values that resolve to another origin.
+- Show an error on the **Agents** page when the list fails to load instead
+  of reporting that there are no agents.
+- Flush heartbeat statistics every 5 seconds instead of every 60, so agent
+  detail pages update sooner.
+- Document date fields, read-only arrays, and nullable values correctly in
+  the OpenAPI document, and align the `GET /api/agents` query parameters
+  with what the API accepts.
+- Keep the caret in place while editing the agent search text.
+
+---
+
+## v1.1.2 {date="2026-10-06"}
 
 > [!Important]
 > #### Heartbeat requests now require Heartbeat write permission
@@ -614,6 +780,6 @@
 
 ---
 
 ## v0.0.1-alpha {date="2026-01-01"}
 
 _Initial alpha build of Telegraf Controller._