Pipelines, Workers - Dropped event metrics, typed Pipelines bindings, and improved setup
Key Points
- Dropped-event metrics and GraphQL dataset
- Schema-specific TypeScript Pipeline bindings
- Simplified wrangler pipelines setup (auto R2/Data Catalog)
Summary
Cloudflare Pipelines shipped three practical improvements: observability for dropped streaming events, generated TypeScript bindings for Pipeline schemas, and a simplified wrangler setup flow that automates R2 and Data Catalog creation.
Key Points
-
Dropped event metrics
- Pipelines now surfaces dropped events (events accepted at ingestion but dropped at delivery) via a dashboard and new metrics.
- You can query dropped-event details programmatically using the pipelinesUserErrorsAdaptiveGroups GraphQL dataset to see counts grouped by errorFamily/errorType.
- Common error types include:
missing_field,type_mismatch,parse_failure, andnull_value.
-
Typed Pipelines bindings
wrangler typesgenerates schema-specific TypeScript types for Pipeline bindings so missing fields and incorrect types are caught at compile time instead of as runtime dropped events.- Update your Worker Env interface to use the generated Pipeline<T> binding to get type safety for records.
-
Improved Pipelines setup
wrangler pipelines setupnow has a Simple setup mode that applies recommended defaults, auto-creates the R2 bucket, and enables R2 Data Catalog if needed.- Validation errors during setup now prompt inline retry rather than forcing a full restart.
Action items for engineers
- Inspect the new dropped-event dashboard and query pipelinesUserErrorsAdaptiveGroups for root-cause analysis over a time range.
- Run
wrangler typesand adopt generated Pipeline<T> types in your Worker Env to catch schema issues at build time. - Use the wrangler pipelines setup Simple mode when creating new pipelines to automate R2 and Data Catalog provisioning and reduce manual steps.
References
- See the Pipelines docs for GraphQL examples, Typed Pipeline bindings, and the Getting started guide for setup details.