Provide an outbound connection from the SDK to the Dev Server and Cloud. Allowing long lived connections to handle long running jobs which can exceed HTTP request timeouts. This also allows users to run the SDK within their network without opening ports.

Adding the "connect" API for long running connections to the Python SDK

A flow control option to ensure that if another run will not be triggered if there is a run that is actively running with the same key.

Allow exporting metrics to Datadog and Prometheus

Allow users to view and verify the function configuration is correct via the UI in the Dev Server.

New events pages in Cloud and Dev Server, with better event search and filter

Allow users to configure alerts for automatic notifications when functions fail

Forward database change events locally to Inngest to trigger functions

Experimental support out now: https://www.inngest.com/docs/self-hosting. Run the open source Inngest on your own infrastructure with minimal dependencies. Postgres support coming soon.

Steps should be able to override retry policies and set their own concurrency keys

Allow passing in regions, eg: eu, within HTTP headers when sending events, ensuring that all data is parsed, stored, and executed within the specified region.

Tools like superjson are common in the JS ecosystem, allowing users to maintain objects such as dates across network boundaries. The SDK can support these patterns by allowing a "transformer" to serialize/deserialize data passing to/from Inngest.

Ability to see webhook requests sent to Inngest. It will help users create and test transform functions and debug webhook issues, such as mismatch of event format.

Forward database change events to Inngest to trigger functions (change data capture)

Improved endpoints for consistency and supporting key developer use cases via the REST API

Add the ability to set a timestamp in the past for waitForEvent to retroactively look backwards for an event. This is useful for race conditions and waitForEvent.

Define concurrency limits on function runs as opposed to individual steps

Improving the DX os webhooks to make it easier to write, test, and debug issues with webhook transforms.

Allow configuration of step timeouts on a per-step basis, cancelling work if the timeout is reached before steps finish.

Write Inngest functions in PHP and serve them using your favorite frameworks like Laravel.

Define triggers (events or crons) via an API rather than in a function definition

Invoke an Inngest function outside of the scope of a function handler and await it's result.

Store function failures and allow full text search (eg. for filtering, which lets us replay specific errors). Categorize errors, eg. for function timeouts, and show smart tips to resolve.

When rerunning a function, start the new run from the failed step

Bulk reprocess past events

Expose low, normal, and high priorities for functions. Potentially allow replays to have their own priorities (note that if a function has high priority and a replay has high priority, this will be a noop).

Add filtering by app, sorting and a filter text input

Add the ability to share throttle configuration across functions with scopes.

New roles for varying levels of access within an Inngest account.

Enable users to select webhooks within their Inngest Cloud account that they want to temporarily forward to a running Dev Server. Useful for testing webhook functions during local dev.

When deploying crons, having them run in each branch environment can make testing more difficult and increase function calls unnecessarily. Users can still manually invoke crons in Branch Envs.

A user testing function that use "sleep" or "sleepUntil" should be able to click a ▶️ or similar button that forces the function to continue and complete the sleep faster. This is key to testing long sleeps.

A pre-build Github action that will trigger an App sync via Inngest REST API. This should provide a more seamless syncing experience similar to the Vercel integration for all users.

Add a tool which allows for continual linear subscriptions to an event stream, eg: step.subscribe("event/name", { if: "xyz == 123", timeout: "10m" }, (acc, async, ctx) => {}). This lets us listen for all events of a type within a function.

Use an SDK to add key-value tags to a given function run enabling future search and filtering.

Users are able to use autocomplete when they type their payloads, based on the schema defined

Improved environments overview page for creating and managing environments including branch environments.

Add ability to manually archive events

Log skipped runs along side events that did not trigger function runs due to flow control configuration like rate limit, debounce, idempotentcy.

Allow webhook transforms to customize the response status code, headers, and body

Forward metrics and other Inngest system data to New Relic.

Clicking a function in the dev server function list should show the configuration options to the user in order to verify their config is correct (e.g. concurrency)

Add CEL expression filtering to bulk cancellation and better integrate into the Cloud dashboard experience.

UX improvements to archived environments

When a user sees a failing run, they

Simpler function configuration, improves types and streamlined developer workflow.

Provide a drop down with typeahead filter global runs views within Cloud + Dev Server