{"componentChunkName":"component---src-components-mark-down-block-js","path":"/docs/cribl-integration/","result":{"data":{"markdownRemark":{"html":"

Integrating with Cribl

\n

By default, AppScope is configured to connect to an instance of Cribl Edge (docs) running on the same host. Cribl Edge, in turn, connects to Cribl Stream over the network. This arrangement can support scaling to large numbers of scoped processes.

\n

AppScope can also easily connect to Cribl Stream (overview | cloud | download | docs) directly.

\n

Integrating with Cribl Edge

\n

AppScope will connect with Cribl Edge when the cribl backend is enabled with the edge transport type. These are default settings for the AppScope library, which is best suited for the kinds of longer-running, planned monitoring procedures that tend to scale, and where the AppScope-Cribl Edge combination excels.

\n

The CLI, by contrast, writes to the local filesystem by default; and the CLI is best suited for ad hoc exploration that's less likely to scale to large numbers of scoped processes.

\n

To define a connection to Cribl Edge with the AppScope library, just set the SCOPE_CRIBL environment variable to edge.

\n

For example:

\n
LD_PRELOAD=./libscope.so SCOPE_CRIBL=edge ls -al\n
\n

Scaling to Many Scoped Processes

\n

The number of scoped processes can become very high in two kinds of situations:

\n
    \n
  1. Where the application that one instance of AppScope is monitoring spawns a large number of processes - for example, scoping a build process; and,
    2. \n
  2. Where many instances of AppScope are monitoring many instances of an application in a distributed architecture - for example, monitoring Apache in a server farm.
    3. \n
\n

In either case, it can help when AppScope connects to Cribl Edge on the same host, and Cribl Edge then connects to Cribl Stream over the network.

\n

To understand why this is, consider these facts:

\n\n

This results in dramatically lower numbers of network connections needed than if AppScope connected to Cribl Stream over the network without Cribl Edge as a proxy. Each network connection requires a port, and a large number of connections could exhaust the supply of ports available on the Cribl Stream host.

\n

Sending data through Cribl Edge has other benefits: Cribl Edge is capable of filtering, aggregating, and reducing data in the same manner as Cribl Stream, and it can be efficient to perform these operations on the data before sending it over the wire.

\n

Integrating with Cribl Stream

\n

To define a TLS-encrypted connection to Cribl Stream on Cribl.Cloud, just set the SCOPE_CRIBL_CLOUD environment variable, specifying a transport type, a host name or IPv4 address, and a port number.

\n

For example:

\n
SCOPE_CRIBL_CLOUD=tcp://in.main-default-<organization>.cribl.cloud:10090\n
\n

By default, Cribl.Cloud-managed instances of Cribl Stream have port 10090 configured to use TLS over TCP, and a built-in AppScope Source to receive data from AppScope. You can change the AppScope Source configuration, or create additional AppScope Sources, as needed.

\n

This is the easiest way to integrate AppScope with Cribl Stream. It is also possible to connect to Cribl Stream on Cribl.Cloud unencrypted; or, to Cribl Stream on-prem, encrypted or not.

\n

Parameter Overrides

\n

When AppScope establishes a Cribl Stream connection, this sets several AppScope configuration parameters, some of which override settings in any configuration file or environment variable.

\n

These overrides include:

\n\n

Any configuration override is logged to AppScope's default log file.

\n

If you have events or metrics enabled (or disabled), they retain this setting. (Before v.1.0.0, both would be set to enabled, as an override). You still have control via environment variables or the config file.

\n

When the payloads feature is enabled, setting SCOPE_PAYLOAD_TO_DISK to true guarantees that AppScope will write payloads to the local directory specified in SCOPE_PAYLOAD_DIR. The payload field in the process-start message records whether the feature is enabled or not, and if enabled, the directory to which it will write payloads.

\n

\n

Connecting to Cribl Stream on Cribl.Cloud, Unencrypted

\n

To define an unencrypted connection to a Cribl.Cloud-managed instance of Cribl Stream, set the SCOPE_CRIBL environment variable (not SCOPE_CRIBL_CLOUD) and specify port 10091.

\n

For example:

\n
SCOPE_CRIBL=tcp://in.main-default-<organization>.cribl.cloud:10091\n
\n

\n

Connecting to Cribl Stream On-Prem

\n

An on-prem instance of Cribl Stream has an AppScope Source built in. However, by default:

\n\n

To connect from AppScope, you'll need to configure the AppScope Source in Cribl Stream. This includes\nenabling the Source; configuring it to listen on the desired port; and, enabling and configuring TLS if desired.

\n

Then, to define a connection to the on-prem Cribl Stream instance, set SCOPE_CRIBL.

\n

For example:

\n
SCOPE_CRIBL=tcp://127.0.0.1:10090\n
","frontmatter":{"title":"Integrating with Cribl"}}},"pageContext":{"slug":"/cribl-integration/","id":"02c5b33d-18ec-5e17-9852-f1cb1bab750c"}},"staticQueryHashes":["2787008656","306729481","625673991","640077088"]}