--- title: "Set up fleet-managed Adaptive Logs | Grafana Cloud documentation" description: "Set up Adaptive Logs drop rules and exemptions on Fleet Management collectors so logs are filtered before they leave your infrastructure." --- > For a curated documentation index, see [llms.txt](/llms.txt). For the complete documentation index, see [llms-full.txt](/llms-full.txt). # Set up fleet-managed Adaptive Logs > Note > > Fleet-managed Adaptive Logs is currently in [public preview](/docs/release-life-cycle/). Grafana Labs offers limited support, and breaking changes might occur prior to the feature being made generally available. Adaptive Logs can apply your [drop rules](../../drop-rules/) and [exemptions](../../create-exemptions/) on collectors you run in your infrastructure, so matching logs are filtered before they leave your environment. Pattern drop rates still apply in Grafana Cloud to the remaining logs. You turn on sync once per tenant in Adaptive Logs. The confirmation tells you that syncing requires setup. You then click **Set up in Fleet Management** and add an attribute to the collectors that should filter logs. After that, manage drop rules and exemptions in Adaptive Logs as usual. You do not author or maintain these pipelines. Adaptive Logs keeps them in sync as your drop rules and exemptions change. Each Grafana Cloud stack gets its own pipelines, which forward with that stack’s credentials. Apps that send logs to the collector on your network cannot pick a different stack. Those logs always go to the stack the pipeline was created for. ## Why filter on your collectors Adaptive Logs tells you which logs you can drop. By default, Grafana Cloud applies those drop rules after the logs have already left your infrastructure. You still pay to send and process volume you intend to drop. Fleet-managed Adaptive Logs applies the same drop rules and exemptions on collectors you already manage in [Grafana Fleet Management](/docs/grafana-cloud/observe-and-act/send-data/fleet-management/). Matching logs are filtered before they leave your environment. You still review, apply, and edit drop rules and exemptions in Adaptive Logs. Adaptive Logs creates the Fleet Management pipelines and keeps them in sync. The one action in Fleet Management is adding an attribute to the collectors that should filter logs. ## Before you begin You need the following: - Administrative privileges for or [role-based access](/docs/grafana-cloud/observe-and-act/send-data/fleet-management/set-up/role-based-access-control/) to Fleet Management on your Grafana Cloud instance - At least one collector in Fleet Management **Inventory**. The empty state offers **Install Alloy**, which enrolls Grafana Alloy with remote configuration. To enroll an OpenTelemetry Collector instead, go to **Connections** > **Collector Setup**, choose **OpenTelemetry**, and install the `otelcol-contrib` distribution. Refer to [Get started with Fleet Management and Grafana Alloy](/docs/grafana-cloud/send-data/fleet-management/get-started/alloy/), [Get started with Fleet Management and OpenTelemetry Collector](/docs/grafana-cloud/send-data/fleet-management/get-started/opentelemetry-collector/), or [onboard existing collectors](/docs/grafana-cloud/send-data/fleet-management/set-up/onboard-collectors/). - Permission to edit Adaptive Logs tenant configuration (`grafana-adaptivelogs-app.tenant-config:edit`). The Adaptive Logs **Admin** role includes this permission. Enroll your first collector to activate Fleet Management. Until a collector appears in **Inventory**, you cannot add the Adaptive Logs attribute, and Adaptive Logs cannot generate pipelines. You never create or edit those pipelines in Fleet Management. After you add the attribute and turn on sync, Adaptive Logs generates the pipelines and sets the matcher. ## Sync pipelines in Adaptive Logs Start in Adaptive Logs. Select the sync checkbox, confirm, then click **Set up in Fleet Management**. Syncing requires that setup; enabling sync alone does not start fleet-managed collection. 1. In Grafana Cloud, navigate to **Administration** > **Plugins and data** > **Plugins** > **Adaptive Logs** > **Configuration**. 2. In **Fleet Management**, marked **Preview**, select **Sync to Fleet Management for this tenant**. 3. In the confirmation dialog box, click **Sync**. The dialog box explains that syncing requires setting up fleet-managed collectors for this tenant. 4. Click **Set up in Fleet Management**. The **Remote configuration** tab opens, filtered to Adaptive Logs pipelines, for example `adaptive_logs__loki_push`, `adaptive_logs__otel`, and `adaptive_logs__otel_collector`. You don’t have to change anything on that screen. Don’t click **Create configuration pipeline**, and don’t add matchers. Sync turns pipeline generation on for this tenant. Adaptive Logs generates pipelines after at least one collector has the Adaptive Logs attribute, and keeps them in sync when you change drop rules or exemptions. If the list is empty, add the collector attribute in the next section, then wait for the next update. To disable a generated pipeline, turn off the **Active** switch and click **Deactivate**. To stop syncing for this tenant, refer to [Turn off fleet-managed Adaptive Logs](#turn-off-fleet-managed-adaptive-logs). ## Add collector attributes in Fleet Management **Set up in Fleet Management** is the follow-on step after you turn on sync. Tag the collectors that should run the Adaptive Logs pipelines. Adaptive Logs does not add the attribute for you. If **Inventory** shows **No collectors found with Fleet Management enabled**, enroll a collector first, then return here after it appears in the list. Click **Install Alloy** for Grafana Alloy. For an OpenTelemetry Collector, use **Connections** > **Collector Setup** and choose **OpenTelemetry**. Don’t use **Install Alloy**. 1. In Fleet Management, open the **Inventory** tab. 2. Select the collectors that should filter logs before they reach Grafana Cloud. 3. Click the edit icon, then **Add remote attributes**. 4. Add the attribute for the collector type: - Grafana Alloy: `adaptive_logs_enabled="true"`. The collector receives both Alloy pipelines, Loki push and Alloy OTLP. You cannot select only one. Unused listeners are harmless if nothing sends to them. - OpenTelemetry Collector: `adaptive_logs_otel_enabled="true"`. The collector receives the OpenTelemetry Collector pipeline. For the general steps, refer to [Add remote attributes](/docs/grafana-cloud/send-data/fleet-management/set-up/onboard-collectors/standalone-alloy/#add-remote-attributes). If you later remove the attribute or clear **Sync to Fleet Management for this tenant**, Fleet Management removes the generated pipelines on the next update. ## Deploy a gateway Run Adaptive Logs on a small set of dedicated collectors that other apps and agents send logs to, not on every per-host agent. Add the Adaptive Logs attribute only to those gateway collectors. If you already send logs through a central collector, opt in that collector and leave per-host agents unchanged. For deployment patterns, refer to [Deploy Grafana Alloy](/docs/grafana-cloud/observe-and-act/send-data/alloy/set-up/deploy/#as-a-centralized-collection-service), [Use Grafana Alloy as a proxy or aggregation layer](/docs/grafana-cloud/observe-and-act/send-data/alloy/configure/proxy/), and [Get started with Fleet Management and OpenTelemetry Collector](/docs/grafana-cloud/send-data/fleet-management/get-started/opentelemetry-collector/). ## Send logs to opted-in collectors Adaptive Logs generates the pipelines. Point log traffic at the collectors you opted in. Expand table | How logs arrive | Collector | Ports | |-----------------|------------------------------------------|------------| | Loki push | Grafana Alloy | 3100 | | OTLP | Grafana Alloy or OpenTelemetry Collector | 4317, 4318 | An Alloy collector that opts in receives both pipelines. Point Loki-push clients at port 3100 and OTLP clients at ports 4317 and 4318. For OpenTelemetry Collector, use the `otelcol-contrib` distribution. > Caution > > The generated receivers accept traffic without authentication. Restrict network access to these ports. ## Turn off fleet-managed Adaptive Logs To stop syncing Adaptive Logs pipelines for this tenant, complete the following steps. 1. In Grafana Cloud, navigate to **Administration** > **Plugins and data** > **Plugins** > **Adaptive Logs** > **Configuration**. 2. In **Fleet Management**, clear **Sync to Fleet Management for this tenant**. 3. Click **Stop syncing**. Fleet Management removes the generated pipelines on the next update. Grafana Cloud applies drop rules and exemptions again to logs that arrive without a collector-processed marker. ## Troubleshoot setup If collectors aren’t filtering logs, or Fleet Management doesn’t show the pipelines you expect, check the following. ### Why is Inventory empty? Until a collector appears in Fleet Management **Inventory**, you can’t add the Adaptive Logs attribute, and Adaptive Logs can’t generate pipelines. Enroll a collector with this stack, then return to **Inventory**. Click **Install Alloy** for Grafana Alloy. For an OpenTelemetry Collector, go to **Connections** > **Collector Setup**, choose **OpenTelemetry**, and install the `otelcol-contrib` distribution. Don’t use **Install Alloy** for OpenTelemetry Collector. ### Why is the Remote configuration list empty? Adaptive Logs generates pipelines only while sync is on and at least one collector has the Adaptive Logs attribute. Add `adaptive_logs_enabled="true"` for Grafana Alloy or `adaptive_logs_otel_enabled="true"` for OpenTelemetry Collector, turn on sync, then wait for the next update. ### Why is the collector failing after the attribute is added? The generated pipelines bind ports 3100, 4317, and 4318 on all interfaces. If the collector already listens on those ports, the listeners conflict. Run fleet-managed pipelines on dedicated collector instances, or make sure nothing else uses these ports. Refer to [Deploy a gateway](#deploy-a-gateway). ### Why are logs still reaching Grafana Cloud? Point log traffic at an opted-in collector. If logs bypass that collector, Grafana Cloud applies drop rules after the logs leave your infrastructure. If a collector can’t evaluate a log, it forwards the log without a processed marker, and Grafana Cloud applies drop rules as usual. Those logs still leave your infrastructure. ### Why has a drop rule or exemption change not taken effect? Pipelines update on an interval. After you change a drop rule or exemption, wait for the next update before you expect opted-in collectors to apply it. If a 100% drop rule has no per-rule drop metrics, refer to [Drop logs before they reach Grafana Cloud](/docs/grafana-cloud/adaptive-telemetry/adaptive-logs/drop-rules/#drop-logs-before-they-reach-grafana-cloud). ## Next steps After collectors are opted in, manage drop rules and exemptions as usual. For the 100% drop-rule alert and **On the edge** chip, refer to [Drop logs before they reach Grafana Cloud](/docs/grafana-cloud/adaptive-telemetry/adaptive-logs/drop-rules/#drop-logs-before-they-reach-grafana-cloud).