Prometheus

Cribl Stream can send metric events to targets and third-party platforms that support Prometheus’ remote write specification (overview here).

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

Configuring Cribl Stream to Output to Prometheus

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 Prometheus. 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 Prometheus. 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 Prometheus output definition.

Remote Write URL: The endpoint to send events to, e.g.: http://localhost:9200/write

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.

Authentication

Select one of the following options for authentication:

  • None: Don’t use authentication.

  • Auth token: Use HTTP token authentication. In the resulting Token field, enter the bearer token that must be included in the HTTP authorization header.

  • Auth token (text secret): This option exposes a Token (text secret) 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: Displays Username and Password fields for you to enter HTTP Basic authentication credentials.

  • 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.

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.

Processing Settings

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. By default, includes cribl_host (Cribl Stream Node that processed the event) and cribl_wp (Cribl Stream Worker Process that processed the event). Supports wildcards. 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:

Status CodeAction
Greater than or equal to 400 and less than or equal to 500.Drop the request.
Greater than 500.Retry the request.

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 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 per Worker Process. When Cribl Stream hits this limit, it begins throttling traffic to the downstream service. Defaults to 5. Minimum: 1. Maximum: 32.

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.

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>.

Send metadata: Whether to generate and send metrics’ metadata (type and metricFamilyName) along with the metrics. The default Yes value displays this additional field:

  • Metadata flush period (sec): How frequently metrics metadata is sent out. Value must at least equal the base Flush period (sec). (In other words, metadata cannot be flushed on a shorter interval.) Defaults to 60 seconds.

Failed request logging mode: Use this drop-down to determine which data should be logged when a request fails. Select among None (the default), Payload, or Payload + Headers. With this last option, Cribl Stream will redact all headers, except non-sensitive headers that you declare below in Safe headers.

Safe headers: Add headers to declare them as safe to log in plaintext. (Sensitive headers such as authorization will always be redacted, even if listed here.) Use a tab or hard return to separate header names.

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 to the HTTP endpoint as a metric event. Otherwise, Cribl Stream will drop the event.

Notes on HTTP-Based Outputs

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

  • Unlike other HTTP-based Destinations, Prometheus does not display an Advanced Settings > Compress option. The Prometheus remote_write spec assumes that payloads are snappy-compressed by default.

  • 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 Prometheus cluster nodes.