Document Executor job function concurrency limit

[dragon3]

Documents the per-workspace concurrency limit on Executor job function operations (100 concurrent executions), which ensures fair scheduling across workspaces.

Changes

• docs/reference/platform/platform-limits.md
  • New row in the Service Limits table (Executor / Workspace Concurrent Job Function Operations / 100).
  • New "Executor Job Function Concurrency Limit" section mirroring the existing Workflow Concurrency Limits structure: how it works, queued-not-rejected behavior, oldest-first ordering, workspace isolation, and a drain-time rule of thumb (queued ÷ 100 × average duration).
  • Clarifies that other operation kinds are not counted against this limit, and that workflow operations are governed by the Workflow concurrency limits on the same page.
  • New Best Practices entry: bulk fan-out patterns (e.g. imports where each record triggers downstream executors) are smoothed to 100 concurrent executions — chain downstream steps on completion events instead of fixed schedules.
• docs/guides/executor/job-function-operation.md
  • New "Concurrency" section summarizing the behavior with a link to the Platform Limits page.

Key behavior documented

Executions above the limit are queued, never rejected or failed; they start automatically, oldest first, as running executions complete. The limit applies to job function operations only, for any trigger type, and each workspace is isolated from other workspaces' bursts.

[dragon3]

Thanks!