This video introduces Cribl Packs, a way to package pipelines and knowledge objects into reusable, deployable units. Packs make it easier to develop and test configurations in one environment and then consistently deploy them across multiple worker groups or production environments.
By abstracting logic away from physical data routing, Packs enable teams to share, reuse, and manage observability configurations at scale. The video also highlights the Cribl Pack Dispensary, which provides pre-built packs to help teams quickly manage common types of observability data.
Full Transcript
This transcript was automatically generated and lightly edited for clarity. Please excuse any transcription errors.
Auto-Generated Transcript (lightly edited for accuracy)
00:06 This video is a walkthrough of Stream Packs. Stream helps people tame their observability data, saving money and making data pipelines more efficient and consistent.
00:14 It enables companies to let teams decide which observability tools are right for their use cases, using the right data.
00:22 But historically, Stream hasn’t known how to share. Want to copy a pipeline from one worker group to another?
00:33 You can copy and paste it, sure—but then how do you maintain both? How do you build pipelines and knowledge objects in a dev environment…
00:41 …and then move them to production when you’re ready? That’s traditionally been a manual and fairly painful process.
00:48 That’s changing with the advent of packs. A pack is a collection of pipelines and knowledge objects bundled together for easy deployment and redeployment.
00:54 Do you have worker groups that are geographically aligned but still need to share pipelines?
01:02 Create a pack of the pipeline configuration and simply deploy that pack to all the worker groups that need it.
01:24 Packs include pipelines and knowledge objects—like lookups or regular expressions—and abstract them from the physical routing of data from source to destination.
01:32 This allows you to bake all of your logic into a reusable pack and then wire it differently for each environment.
01:39 We’ve also created the Cribl Pack Dispensary, a curated set of packs you can use to manage common types of observability data.
01:47 You can install packs by uploading a .crbl file, pointing to one by URL, or referencing a Git repository that contains a pack.
02:00 Pack once, stream everywhere. Thanks for watching.