initializing ErrorEventsThreshold

[Brett Samblanet (brettsam)]

Resolves #11871

Pull request checklist

IMPORTANT: Currently, changes must be backported to the in-proc branch to be included in Core Tools and non-Flex deployments.

• Backporting to the in-proc branch is not required
  • Otherwise: Backport PR: TBD
• My changes do not require documentation changes
• My changes should not be added to the release notes for the next release
  • Otherwise: I've added my notes to release_notes.md
• My changes do not need to be backported to a previous version
• My changes do not require diagnostic events changes
• I have added all required tests (Unit tests, E2E tests)

Additional information

During worker indexing, a webhost language worker can publish a startup error before RpcFunctionInvocationDispatcher.InitializeAsync receives function metadata. The dispatcher subscribed to worker errors in its constructor, but its restart threshold was not initialized until after the empty-function early return. This left the threshold at 0, causing the first startup error to immediately exhaust the retry budget and stop the host.

This change:

• Defaults the restart threshold to 3 and initializes it from the configured worker process count before subscribing to worker events.
• Retains the existing post-metadata recalculation for runtimes resolved from function metadata.
• Avoids setting up invocation buffers or sending function-load requests when a replacement worker starts before metadata is available.

The regression test covers worker errors both before dispatcher initialization and after empty metadata initialization. Before the fix, it failed with Expected: RestartWorker / Actual: StopApplication; after the fix, the configured retry budget is consumed and worker recovery is attempted. Additional coverage verifies the default threshold remains at least 3.

[Copilot]

Pull request overview

Adjusts RpcFunctionInvocationDispatcher to ensure ErrorEventsThreshold is initialized early (with a safe default and pre-event-subscription initialization) so worker startup/indexing errors don’t immediately exhaust the retry budget and stop the host, addressing #11871.

Changes:

• Default ErrorEventsThreshold to 3 and initialize it earlier based on configured process count when possible.
• Recalculate the threshold after function metadata initialization using the resolved max process count.
• Avoid setting up invocation buffers / sending load requests when starting replacement workers before metadata is available, and add regression tests for the indexing-first-error scenario.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

File	Description
test/WebJobs.Script.Tests/Workers/Rpc/RpcFunctionInvocationDispatcherTests.cs	Adds coverage for default threshold behavior and worker-indexing error recovery before/after empty-metadata initialization.
src/Functions.Rpc.Server/Rpc/FunctionRegistration/RpcFunctionInvocationDispatcher.cs	Initializes ErrorEventsThreshold earlier and skips buffer/load setup when _functions is not yet available.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[Brett Samblanet (brettsam)]

Ok this is more complicated than I first assumed. We've never tried to restart the process in this state and it's caused all kind of issues. Converting to draft to do more investigation.