Grafana Cloud

Cribl Stream can send data to two of the services available in Grafana Cloud: Loki for logs and Prometheus for metrics. The Grafana Cloud Destination shapes events appropriately for Loki and Prometheus, and routes events to the correct endpoint for each service.

Type: Streaming | TLS Support: Configurable | PQ Support: Yes

Preparing Prometheus and Loki to Receive Data from Cribl Stream​

To define a Grafana Cloud Destination, you need a Grafana Cloud account.

While logged in to your Grafana account, navigate to the Grafana Cloud Portal, which should be located at https://grafana.com/orgs/<your-organization-name>, and complete the following steps.

Obtain an API key, setting its Role to MetricsPublisher. If you want Cribl Stream or an external KMS to manage the API key, configure a key pair that references the API key.

In the Prometheus tile, click Send Metrics to open the Prometheus configuration page. Write down:

• Your Remote Write Endpoint URL, for example:
  https://prometheus-blocks-prod-us-central1.grafana.net/api/prom/push.
• Your Prometheus Username.

In the Loki tile, click Send Logs to open the Loki configuration page. Write down:

• Your Grafana Data Source settings URL, for example:
  https://logs-prod-us-central1.grafana.net.
• Your Loki User ID.

Decide what type of authentication to use and prepare accordingly:

• If you choose Basic authentication, the username (Username in Prometheus, User in Loki) and password (simply your Grafana API key) will remain separate.

• If you choose token-based authentication, construct your tokens by concatenating username, colon (:), and password, for example 12345:cOQvDj6sJGFS3Bk2MguBW==. Because the Prometheus and Loki usernames differ, you need to construct a separate token for each service.

Configuring Cribl Stream to Output to Grafana Cloud​

From the top nav, click Manage, then select a Worker Group to configure. Next, you have two options:

To configure via the graphical QuickConnect UI, click Routing > QuickConnect (Stream) or Collect (Edge). Next, click Add Destination at right. From the resulting drawer’s tiles, select Grafana Cloud. Next, click either Add Destination or (if displayed) Select Existing. The resulting drawer will provide the options below.

Or, to configure via the Routing UI, click Data > Destinations (Stream) or More > Destinations (Edge). From the resulting page’s tiles or the Destinations left nav, select Grafana Cloud. Next, click Add Destination to open a New Destination modal that provides the options below.

General Settings​

Output ID: Enter a unique name to identify this Grafana Cloud output definition. If you clone this Destination, Cribl Stream will add -CLONE to the original Output ID.

Loki URL: The endpoint to send log events to, e.g.: https://logs-prod-us-central1.grafana.net. This is the Grafana Data Source settings URL you wrote down earlier.

Prometheus URL: The endpoint to send metric events to, e.g.:
https://prometheus-blocks-prod-us-central1.grafana.net/api/prom/push. This is the Remote Write Endpoint URL you wrote down earlier.

Optional Settings​

Backpressure behavior: Whether to block, drop, or queue events when all receivers are exerting backpressure.

Tags: Optionally, add tags that you can use to filter and group Destinations in Cribl Stream’s Manage Destinations page. These tags aren’t added to processed events. Use a tab or hard return between (arbitrary) tag names.

Persistent Queue Settings​

This tab is displayed when the Backpressure behavior is set to Persistent Queue.

On Cribl-managed Cribl.Cloud Workers (with an Enterprise plan), this tab exposes only the destructive Clear Persistent Queue button (described below in this section). A maximum queue size of 1 GB disk space is automatically allocated per PQ‑enabled Destination, per Worker Process. The 1 GB limit is on outbound uncompressed data, and no compression is applied to the queue.

This limit is not configurable. If the queue fills up, Cribl Stream will block outbound data. To configure the queue size, compression, queue-full fallback behavior, and other options below, use a hybrid Group.

Max file size: The maximum data volume to store in each queue file before closing it. Enter a numeral with units of KB, MB, etc. Defaults to 1 MB.

Max queue size: The maximum amount of disk space the queue is allowed to consume. Once this limit is reached, queueing is stopped and data blocking is applied. Enter a numeral with units of KB, MB, etc.

Queue file path: The location for the persistent queue files. This will be of the form: your/path/here/<worker-id>/<output-id>. Defaults to: $CRIBL_HOME/state/queues.

Compression: Codec to use to compress the persisted data, once a file is closed. Defaults to None. Gzip is also available.

Queue-full behavior: Whether to block or drop events when the queue is exerting backpressure (because disk is low or at full capacity). Block is the same behavior as non-PQ blocking, corresponding to the Block option on the Backpressure behavior drop-down. Drop new data throws away incoming data, while leaving the contents of the PQ unchanged.

Clear Persistent Queue: Click this “panic” button if you want to delete the files that are currently queued for delivery to this Destination. A confirmation modal will appear - because this will free up disk space by permanently deleting the queued data, without delivering it to downstream receivers. (Appears only after Output ID has been defined.)

Strict ordering: The default Yes position enables FIFO (first in, first out) event forwarding. When receivers recover, Cribl Stream will send earlier queued events before forwarding newly arrived events. To instead prioritize new events before draining the queue, toggle this off. Doing so will expose this additional control:

• Drain rate limit (EPS): Optionally, set a throttling rate (in events per second) on writing from the queue to receivers. (The default 0 value disables throttling.) Throttling the queue’s drain rate can boost the throughput of new/active connections, by reserving more resources for them. You can further optimize Workers’ startup connections and CPU load at Group Settings > Worker Processes.

Authentication​

The Authentication tab provides separate Loki and Prometheus sections, enabling you to configure these inputs separately. The two sections provide identical options.

Select one of the following options for authentication:

• Auth token: Enter the bearer token that must be included in the authorization header. Use the token that you constructed earlier. In Grafana Cloud, the bearer token is generally built by concatenating the username and the API key, separated by a colon. E.g.: <your-username>:<your-api-key>.

• Auth token (text secret): This option exposes a drop-down in which you can select a stored text secret that references the bearer token described above. A Create link is available to store a new, reusable secret.

• Basic: This default option displays fields for you to enter HTTP Basic authentication credentials. Username is the Loki User or Prometheus Username that you wrote down earlier. Password is your API key in the Grafana Cloud domain.

• Basic (credentials secret): This option exposes a Credentials secret drop-down, in which you can select a stored text secret that references the Basic authentication credentials described above. A Create link is available to store a new, reusable secret.

Processing Settings​

Metric events can have dimensions, and log events have labels. Dimensions, labels, and their values are determined by several different settings in Cribl Stream. This section explains how that works, along with other kinds of settings.

Loki uses labels to define separate streams of logging data. This is a key concept. Cribl recommends that you familiarize yourself with the information and documentation Grafana provides about labels in Loki.

One canonical example is processing logs from servers in three environments: production, staging, and testing. You could create a label named env whose possible values are prod, staging, and test.

One basic principle is that if you set too many labels, you can end up with too many streams.

Post‑Processing​

Pipeline: Pipeline to process data before sending the data out using this output.

System fields: A list of fields to automatically add to events that use this output—both metric events, as dimensions; and, log events, as labels. Supports wildcards.

By default, includes cribl_host (Cribl Stream Node that processed the event) and cribl_wp (Cribl Stream Worker Process that processed the event). On the Loki side, this creates different streams, which prevents Loki from rejecting some events as being out of order when different Nodes or Worker Processes are emitting at different rates.

Other options include:

• cribl_input – Cribl Stream Source that processed the event.
• cribl_output – Cribl Stream Destination that processed the event.
• cribl_pipe – Cribl Stream Pipeline that processed the event.
• cribl_route – Cribl Stream Route (or QuickConnect) that processed the event.

Retries​

Honor Retry-After header: Whether to honor a Retry-After header, provided that the header specifies a delay no longer than 180 seconds. Cribl Stream limits the delay to 180 seconds even if the Retry-After header specifies a longer delay. When enabled, any Retry-After header received takes precedence over all other options configured in the Retries section. When disabled, all Retry-After headers are ignored.

Settings for failed HTTP requests: When you want to automatically retry requests that receive particular HTTP response status codes, use these settings to list those response codes.

For any HTTP response status codes that are not explicitly configured for retries, Cribl Stream applies the following rules:

Upon receiving a response code that’s on the list, Cribl Stream first waits for a set time interval called the Pre-backoff interval and then begins retrying the request. Time between retries increases based on an exponential backoff algorithm whose base is the Backoff multiplier, until the backoff multiplier reaches the Backoff limit (ms). At that point, Cribl Stream continues retrying the request without increasing the time between retries any further.

By default, this Destination has no response codes configured for automatic retries. For each response code you want to add to the list, click Add Setting and configure the following settings:

• HTTP status code: A response code that indicates a failed request, for example 429 (Too Many Requests) or 503 (Service Unavailable).
• Pre-backoff interval (ms): The amount of time to wait before beginning retries, in milliseconds. Defaults to 1000 (one second).
• Backoff multiplier: The base for the exponential backoff algorithm. A value of 2 (the default) means that Cribl Stream will retry after 2 seconds, then 4 seconds, then 8 seconds, and so on.
• Backoff limit (ms): The maximum backoff interval Cribl Stream should apply for its final retry, in milliseconds. Default (and minimum) is 10,000 (10 seconds); maximum is 180,000 (180 seconds, or 3 minutes).

Retry timed-out HTTP requests: When you want to automatically retry requests that have timed out, toggle this control on to display the following settings for configuring retry behavior:

• Pre-backoff interval (ms): The amount of time to wait before beginning retries, in milliseconds. Defaults to 1000 (one second).
• Backoff multiplier: The base for the exponential backoff algorithm. A value of 2 (the default) means that Cribl Stream will retry after 2 seconds, then 4 seconds, then 8 seconds, and so on.
• Backoff limit (ms): The maximum backoff interval Cribl Stream should apply for its final retry, in milliseconds. Default (and minimum) is 10,000 (10 seconds); maximum is 180,000 (180 seconds, or 3 minutes).

Advanced Settings ​

Validate server certs: Reject certificates that are not authorized by a CA in the CA certificate path, or by another trusted CA (e.g., the system’s CA). Defaults to Yes.

Round-robin DNS: Toggle to Yes to use round-robin DNS lookup across multiple IPv6 addresses. When a DNS server returns multiple addresses, this will cause Cribl Stream to cycle through them in the order returned.

Request timeout: Amount of time (in seconds) to wait for a request to complete before aborting it. Defaults to 30.

Request concurrency: Maximum number of concurrent requests before blocking. This is set per Worker Process. Defaults to 5.

Max body size (KB): Maximum size of the request body before compression. Defaults to 4096 KB. The actual request body size might exceed the specified value because the Destination adds bytes when it writes to the downstream receiver. Cribl recommends that you experiment with the Max body size value until downstream receivers reliably accept all events.

Max events per request: Maximum number of events to include in the request body. The 0 default allows unlimited events.

Loki and Prometheus might complain about entries being delivered out of order when Request concurrency is set > 1 and any of Flush period (sec), Max body size (KB), or Max events per request are set to low values.

Flush period (sec): Maximum time between requests. Low values could cause the payload size to be smaller than its configured maximum. Defaults to 1.

Extra HTTP headers: Name-value pairs to pass as additional HTTP headers. Values will be sent encrypted.

Metric renaming expression: A JavaScript expression that can be used to rename metrics. The default expression – name.replace(/\\./g, \'_\') – replaces all . characters in a metric’s name with the Prometheus-supported _ character. Use the name global variable to access the metric’s name. You can access event fields’ values via __e.<fieldName>.

Message format: Whether to send events as Protobuf (the default) or JSON.

Compress: When the Message format is JSON, this setting controls the data compression format used before sending the data to Grafana Cloud. Defaults to Yes for GZIP-compression. (Applies only to Loki’s JSON payloads. This toggle is hidden when the Message format is Protobuf, because both Prometheus’ and Loki’s Protobuf implementations are Snappy-compressed by default.)

Logs message field: The event field to send as log output, for example: _raw. All other event fields are discarded. If left blank, Cribl Stream sends a JSON representation of the whole event.

Logs labels: Name/value pairs where the value can be a static or dynamic expression that has access to all log event fields.

Failed request logging mode: Determines which data is logged when a request fails. Use the drop-down to select one of these options:

• None (default).
• Payload.
• Payload + Headers. Use the Safe Headers field below to specify the headers to log. If you leave that field empty, all headers are redacted, even with this setting.

Safe headers: List the headers you want to log, in plain text.

Environment: If you’re using GitOps, optionally use this field to specify a single Git branch on which to enable this configuration. If empty, the config will be enabled everywhere.

Internal Fields​

Cribl Stream uses a set of internal fields to assist in forwarding data to a Destination.

If an event contains the internal field __criblMetrics, Cribl Stream will send it Prometheus as a metric event. If __criblMetrics is absent, Cribl Stream will treat the event as a log and send it to Loki.

The internal field __labels specifies labels to add to log events. If a label is set in both the __labels field and in Logs labels and/or System fields, Cribl Stream sends the value from __labels to Loki. Setting the __labels field in a Pipeline gives you a quick way to experiment with the logs being sent.

If there are no labels set (this would happen when System fields, Logs labels, and __labels are all empty), Cribl Stream adds a default source label, which prevents Loki from rejecting events. The source label the concatenation of cribl, underscore (_), source type, colon (:), source-name, where source name and type are values in the __inputId event field, for example: cribl_metrics:in_prometheus_rw. If __inputId is missing, source is set to cribl.

Notes on HTTP-Based Outputs​

• To proxy outbound HTTP/S requests, see System Proxy Configuration.

• The Advanced Settings > Compress toggle determines whether to compress the payload body before sending to Loki only. The toggle setting does not apply to Prometheus payloads, which are always compressed using Snappy.

• Cribl Stream will attempt to use keepalives to reuse a connection for multiple requests. After two minutes of the first use, the connection will be thrown away, and a new one will be reattempted. This is to prevent sticking to a particular destination when there is a constant flow of events.

• If the server does not support keepalives (or if the server closes a pooled connection while idle), a new connection will be established for the next request.

• When resolving the Destination’s hostname, Cribl Stream will pick the first IP in the list for use in the next connection. Enable Round-robin DNS to better balance distribution of events between Grafana Cloud nodes.