On This Page

Home / Stream/ Integrations/ Sources/ Prometheus/Prometheus Remote Write Source

Prometheus Remote Write Source

Cribl Stream supports receiving metric data from Prometheus instances that are configured to send data via the remote write protocol.

Type: Push | TLS Support: YES | Event Breaker Support: No

This Source assumes that incoming data is snappy-compressed.

Configure Cribl Stream to Receive Metrics from Prometheus Remote Write Sources
​

  1. On the top bar, select Products, and then select Cribl Stream. Under Worker Groups, select a Worker Group. Next, you have two options:
    • To configure via QuickConnect, navigate to Routing > QuickConnect (Stream) or Collect (Edge). Select Add Source and select the Source you want from the list, choosing either Select Existing or Add New.
    • To configure via the Routes, select Data > Sources (Stream) or More > Sources (Edge). Select the Source you want. Next, select Add Source.
  2. In the New Source modal, configure the following under General Settings:
    • Enabled: Toggle on to enable the Source.
    • Input ID: Enter a unique name to identify this Source definition. If you clone this Source, Cribl Stream will add -CLONE to the original Input ID.
    • Description: Optionally, enter a description.
    • Address: Enter the hostname/IP to listen to. Defaults to 0.0.0.0.
    • Port: Enter the port number to listen on.
    • Remote Write API endpoint: Enter the absolute path on which to listen for Prometheus requests. Defaults to /write, which will (in this example) expand as: http://<your‑upstream‑URL>:<your‑port>/write.
  3. Next, you can configure the following Optional Settings:
    • Tags: Optionally, add tags that you can use to filter and group Sources in Cribl Stream’s UI. These tags aren’t added to processed events. Use a tab or hard return between (arbitrary) tag names.
  4. Optionally, you can adjust the Authentication, TLS, Persistent Queue Settings, Processing, and Advanced settings, or Connected Destinations outlined in the sections below.
  5. Select Save, then Commit & Deploy.

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, or select Generate if you need a new token.

  • Auth token (text secret): Provide an HTTP token referenced by a secret. Select a stored text secret in the resulting drop-down, or select Create to configure a new secret.

  • Basic: Displays Username and Password fields for you to enter HTTP Basic authentication credentials. Click Generate if you need a new password.

  • Basic (credentials secret): Provide username and password credentials referenced by a secret. Select a stored text secret in the resulting Credentials secret drop-down, or select Create to configure a new secret.

TLS Settings (Server Side)
​

Enabled: Defaults to toggled off. When toggled on:

Certificate name: Name of the predefined certificate.

Private key path: Server path containing the private key (in PEM format) to use. Path can reference $ENV_VARS.

Passphrase: Passphrase to use to decrypt private key.

Certificate path: Server path containing certificates (in PEM format) to use. Path can reference $ENV_VARS.

CA certificate path: Server path containing CA certificates (in PEM format) to use. Path can reference $ENV_VARS.

Authenticate client (mutual auth): Require clients to present their certificates. Used to perform mutual authentication using SSL certs. Default is toggled off. When toggled on:

  • Validate client certs: Reject certificates that are not authorized by a CA in the CA certificate path, or by another trusted CA (for example, the system’s CA). Default is toggled off.

  • Common Name: Regex that a peer certificate’s subject attribute must match in order to connect. Defaults to .*. Matches on the substring after CN=. As needed, escape regex tokens to match literal characters. (For example, to match the subject CN=worker.cribl.local, you would enter: worker\.cribl\.local.) If the subject attribute contains Subject Alternative Name (SAN) entries, the Source will check the regex against all of those but ignore the Common Name (CN) entry (if any). If the certificate has no SAN extension, the Source will check the regex against the single name in the CN.

Minimum TLS version: Optionally, select the minimum TLS version to accept from connections.

Maximum TLS version: Optionally, select the maximum TLS version to accept from connections.

Persistent Queue Settings
​

In the Persistent Queue Settings tab, you can optionally specify persistent queue storage, using the following controls. Persistent queue buffers and preserves incoming events when a downstream Destination has an outage or experiences backpressure.

Before enabling persistent queue, learn more about persistent queue behavior and how to optimize it with your system:

On Cribl-managed Cloud Workers (with an Enterprise plan), this tab exposes only the Enable Persistent Queue toggle. If enabled, PQ is automatically configured in Always On mode, with a maximum queue size of 1 GB disk space allocated per PQ‑enabled Source, per Worker Process.

The 1 GB limit is on uncompressed inbound data, and the queue does not perform any compression. This limit is not configurable. For configurable queue size, compression, mode, and other options below, use a hybrid Group.

Enable Persistent Queue: Default is toggled off. When toggled on:

Mode: Select a condition for engaging persistent queues.

  • Always On: This default option will always write events to the persistent queue, before forwarding them to the Cribl Stream data processing engine.
  • Smart: This option will engage PQ only when the Source detects backpressure from the Cribl Stream data processing engine.

Smart mode only engages when necessary, such as when a downstream Destination becomes blocked and the Max buffer size reaches its limit. When persistent queue is set to Smart mode, Cribl attempts to flush the queue when every new event arrives. The only time events stay in the buffer is when a downstream Destination becomes blocked.

Max buffer size: The maximum number of events to hold in memory before reporting backpressure to the sender and writing the queue to disk. Defaults to 1000. This buffer is for all connections, not just per Worker Process. For that reason, this can dramatically expand memory usage. Connections share this limit, which may result in slightly lower throughput for higher numbers of connections. For higher numbers of connections, consider increasing the limit.

Commit frequency: The number of events to send downstream before committing that Stream has read them. Defaults to 42.

Max file size: The maximum data volume to store in each queue file before closing it and (optionally) applying the configured Compression. Enter a numeral with units of KB, MB, and so forth. If not specified, Cribl Stream applies the default 1 MB.

Max queue size: The maximum amount of disk space that the queue is allowed to consume on each Worker Process. Once this limit is reached, this Source will stop queueing data and block incoming data. Required, and defaults to 5 GB. Accepts positive numbers with units of KB, MB, GB, and so forth. Can be set as high as 1 TB, unless you’ve configured a different Worker Process PQ size limit in Group/Fleet settings.

Queue file path: The location for the persistent queue files. Defaults to $CRIBL_HOME/state/queues. To this field’s specified path, Cribl Stream will append /<worker-id>/inputs/<input-id>.

Compression: Optional codec to compress the persisted data after a file closes. Defaults to None; Gzip is also available.

In Cribl Stream 4.1 and later, the Source persistent queue default Mode is Always on, to best ensure events’ delivery. For details on optimizing this selection, see Optimize Source Persistent Queues (sPQ).

You can optimize Workers’ startup connections and CPU load at Group/Fleet settings > Worker Processes.

Processing Settings
​

Fields
​

In this section, you can add Fields to each event using Eval-like functionality.

Name: Field name.

Value: JavaScript expression to compute field’s value, enclosed in quotes or backticks. (Can evaluate to a constant.)

Pre-Processing
​

In this section’s Pipeline drop-down list, you can select a single existing Pipeline or Pack to process data from this input before the data is sent through the Routes.

Advanced Settings
​

Show originating IP: Toggle on when clients are connecting through a proxy that supports the X-Forwarded-For header to keep the client’s original IP address on the event instead of the proxy’s IP address. This setting affects how the Source handles the __srcIpPort field.

Capture request headers: Toggle on to add request headers to events, in the __headers field.

Health check endpoint: Toggle on to enable a health check endpoint specific to this Source, http(s)://<host>:<port>/cribl_health. A 200 HTTP response code is returned when the Source is healthy. Otherwise, two errors you could receive are:

  • ECONNRESET where the Source failed to initialize due to not having listeners on the port.
  • 503 or Server is busy, max active connections reached indicate there are too many connections per Worker Process.

Max active requests: Maximum number of active requests allowed for this Source, per Worker Process. Defaults to 256. Enter 0 for unlimited.

Activity log sample rate: Determines how often request activity is logged at the info level. The default 100 value logs every 100th value; a 1 value would log every request; a 10 value would log every 10th request; etc.

Max requests per socket: The maximum number of requests Cribl Stream should allow on one socket before instructing the client to close the connection. Defaults to 0 (unlimited). See Balancing Connection Reuse Against Request Distribution below.

Socket timeout (seconds): How long Cribl Stream should wait before assuming that an inactive socket has timed out. The default 0 value means wait forever.

Request timeout (seconds): How long to wait for an incoming request to complete before aborting it. The default 0 value means wait indefinitely.

Keep-alive timeout (seconds): After the last response is sent, Cribl Stream will wait this long for additional data before closing the socket connection. Defaults to 5 seconds; minimum is 1 second; maximum is 600 seconds (10 minutes).

The longer the Keep‑alive timeout, the more Cribl Stream will reuse connections. The shorter the timeout, the closer Cribl Stream gets to creating a new connection for every request. When request frequency is high, you can use longer timeouts to reduce the number of connections created, which mitigates the associated cost.

IP allowlist regex: Grants access to requests originating from specific IP addresses that match a defined pattern. Unmatched requests are rejected with a 403 (Forbidden) status code. Defaults to .* (allow all).

IP denylist regex: Blocks requests originating from specific IP addresses that match a defined pattern, even if they would be allowed by default. Rejected requests receive a 403 (Forbidden) status code. Defaults to ^$ (allow all).

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.

Balancing Connection Reuse Against Request Distribution
​

Max requests per socket allows you to limit the number of HTTP requests an upstream client can send on one network connection. Once the limit is reached, Cribl Stream uses HTTP headers to inform the client that it must establish a new connection to send any more requests. (Specifically, Cribl Stream sets the HTTP Connection header to close.) After that, if the client disregards what Cribl Stream has asked it to do and tries to send another HTTP request over the existing connection, Cribl Stream will respond with an HTTP status code of 503 Service Unavailable.

Use this setting to strike a balance between connection reuse by the client, and distribution of requests among one or more Worker Node processes by Cribl Stream:

  • When a client sends a sequence of requests on the same connection, that is called connection reuse. Because connection reuse benefits client performance by avoiding the overhead of creating new connections, clients have an incentive to maximize connection reuse.

  • Meanwhile, a single process on that Worker Node will handle all the requests of a single network connection, for the lifetime of the connection. When receiving a large overall set of data, Cribl Stream performs better when the workload is distributed across multiple Worker Node processes. In that situation, it makes sense to limit connection reuse.

There is no one-size-fits-all solution, because of variation in the size of the payload a client sends with a request and in the number of requests a client wants to send in one sequence. Start by estimating how long connections will stay open. To do this, multiply the typical time that requests take to process (based on payload size) times the number of requests the client typically wants to send.

If the result is 60 seconds or longer, set Max requests per socket to force the client to create a new connection sooner. This way, more data can be spread over more Worker Node processes within a given unit of time.

For example: Suppose a client tries to send thousands of requests over a very few connections that stay open for hours on end. By setting a relatively low Max requests per socket, you can ensure that the same work is done over more, shorter-lived connections distributed between more Worker Node processes, yielding better performance from Cribl Stream.

A final point to consider is that one Cribl Stream Source can receive requests from more than one client, making it more complicated to determine an optimal value for Max requests per socket.

Connected Destinations
​

Select Send to Routes to enable conditional routing, filtering, and cloning of this Source’s data via the Routing table.

Select QuickConnect to send this Source’s data to one or more Destinations via independent, direct connections.

Internal Fields
​

Cribl Stream uses a set of internal fields to assist in handling of data. These “meta” fields are not part of an event, but they are accessible, and Functions can use them to make processing decisions.

Fields for this Source:

  • __criblMetrics
  • __final
  • __headers – Added only when Advanced Settings > Capture request headers is toggled on.
  • __inputId
  • __srcIpPort – See details below.
  • _time
  • _value

Overriding __srcIpPort with Client IP/Port
​

The __srcIpPort field’s value contains the IP address and (optionally) port of the Prometheus Remote Write client sending data to this Source.

When any proxies (including load balancers) lie between the Prometheus Remote Write client and the Source, the last proxy adds an X‑Forwarded‑For header whose value is the IP/port of the original client. With multiple proxies, this header’s value will be an array, whose first item is the original client IP/port.

If X‑Forwarded‑For is present, and Advanced Settings > Show originating IP is toggled off, the original client IP/port in this header will override the value of __srcIpPort.

If Show originating IP is toggled on, the X‑Forwarded‑For header’s contents will not override the __srcIpPort value. (Here, the upstream proxy can convey the client IP/port without using this header.)

Detecting Metrics’ Types
​

Because Prometheus remote write requests don’t specify metrics’ types, Cribl Stream applies the following rules to determine the type as we ingest them:

  • If the metric’s name ends with _total, _sum, _count, or _bucket, the type is set to counter.

  • Otherwise, the metric’s type is set to gauge.

This is consistent with the type detection practiced by other services implementing the remote write protocol. See, for example, New Relic’s and Elastic’s documentation.

Note that Cribl Stream supports the timer type in addition to counter and gauge.

Troubleshooting
​

The Source’s configuration modal has helpful tabs for troubleshooting:

Live Data: Try capturing live data to see real-time events as they are ingested. On the Live Data tab, click Start Capture to begin viewing real-time data.

Logs: Review and search the logs that provide detailed information about the ingestion process, including any errors or warnings that may have occurred.

You can also view the Monitoring page that provides a comprehensive overview of data volume and rate, helping you identify ingestion issues. Analyze the graphs showing events and bytes in/out over time.

Common Issue
​

Dropping request because token invalid",“authToken”: “Bas…Njc=”
​

The specified token is invalid. Note that the above message is logged only at the debug level.