Azure Logic Apps Standard – SharePoint Polling Trigger Concurrency, Maximum Waiting Runs, and Trigger Behavior When Queue Is Full

Vetrivel Murugan 0 Reputation points
2026-11-20T06:45:18.18+00:00

Hi Microsoft Azure Team,

We are using Azure Logic Apps Standard Plan with a SharePoint trigger.

Current workflow

  • We have a SharePoint trigger that detects when an item is deleted.
  • After the trigger fires, we retrieve the SharePoint item metadata.
  • We then update the corresponding record in a SQL table.
  • When multiple SharePoint items are deleted simultaneously, multiple workflow runs execute concurrently, resulting in SQL deadlock issues.

Currently, trigger concurrency control is disabled.

Proposed solution

To avoid SQL deadlocks, we are planning to enable trigger concurrency control and process workflow runs sequentially.

Our proposed configuration:

json

"runtimeConfiguration": {
  "concurrency": {
    "runs": 1,
    "maximumWaitingRuns": 100
  }
}

Expected scenario

The SharePoint trigger is a polling trigger that checks for deleted items every minute.

Suppose 5,000 SharePoint items are deleted.

Our understanding of the proposed configuration is:

  • 1 workflow run – Running
  • 100 workflow runs – Waiting in queue
  • Remaining 4,899 items – Pending

Questions

  1. When the 100 maximum waiting runs limit is reached, what happens to the remaining SharePoint trigger events?
  2. Will the SharePoint polling trigger automatically stop polling while the waiting runs queue is full?
  3. Once an existing workflow run completes and queue capacity becomes available, will polling automatically resume and pick up the remaining deleted SharePoint items?
  4. Will the remaining events receive HTTP 429 (Too Many Requests), or will Azure Logic Apps manage polling automatically?
  5. Is there any possibility of missing SharePoint deletion events when the waiting runs queue reaches its maximum capacity?
  6. Is maximumWaitingRuns: 100 a recommended value for this scenario, or should we use a different value for an Azure Logic Apps Standard workflow?

Generated image 2

Expected architecture

We want to ensure that all 5,000 SharePoint deletion events are processed reliably, sequentially, without SQL deadlocks or lost events.

Could Microsoft please clarify the expected trigger behavior and recommend the appropriate concurrency and maximum waiting runs configuration for this scenario?

Thank you.

Azure Logic Apps
Azure Logic Apps

An Azure service that automates the access and use of data across clouds without writing code.

0 comments No comments

1 answer

Sort by: Oldest
  1. AI answer

    2026-11-20T06:46:29.4933333+00:00

    Enable trigger concurrency with runs: 1, but do not treat maximumWaitingRuns as a durable queue or rely on the SharePoint polling trigger alone to guarantee delivery of a 5,000-deletion burst.

    "runtimeConfiguration": {
      "concurrency": {
        "runs": 1,
        "maximumWaitingRuns": 100
      }
    }
    

    This configuration permits one active workflow run and up to 100 waiting runs. Setting runs to 1 is the documented way to process trigger-created workflow instances sequentially.^1^

    Behavior when the waiting queue is full

    For a polling trigger:

    1. The active run continues.
    2. Already-created runs remain in the waiting queue.
    3. When that queue reaches its limit, Logic Apps stops making polling attempts.
    4. After run capacity becomes available, scheduled polling can continue and the trigger can query SharePoint again.^1^

    A full waiting queue does not normally send HTTP 429 responses to the SharePoint deletion events. Logic Apps returns 429 when a request or webhook trigger receives an inbound request that cannot be admitted; a polling trigger instead skips or stops polling while full.^1^

    That is separate from SharePoint connector throttling. SharePoint can return 429 when connector API calls exceed its service limits; the documented connector limit is 600 API calls per connection per 60 seconds.^2^

    The 5,000-event model is not exact

    The events cannot be assumed to divide into exactly:

    • 1 running
    • 100 waiting
    • 4,899 durably pending in SharePoint

    The SharePoint trigger can return an array and use splitOn to create separate workflow runs. Enabling trigger concurrency reduces the splitOn debatching limit to 100 items. If a returned array exceeds that limit, splitOn is disabled.^3^

    More importantly, the available SharePoint connector documentation does not specify a durable delivery contract for deleted-item events when Logic Apps suspends polling for an extended period. Polling triggers must maintain state to avoid rereading data, but whether unpolled events remain available depends on the source and connector implementation.^4^

    Consequently, the waiting-run limit by itself is not sufficient to claim that all 5,000 deletion events will eventually be delivered without loss.

    Separate event capture from sequential SQL processing:

    1. SharePoint ingestion workflow
      • Poll for deleted items.
      • Perform only the minimum processing needed to construct a message.
      • Write each deletion record, including a stable deletion/item identifier, to an Azure Service Bus queue.
      • Do not perform the SQL update in this workflow.
    2. Service Bus processing workflow
      • Trigger from the queue in peek-lock mode.
      • Set concurrency to 1 if every SQL update must be globally serialized.
      • Update SQL.
      • Complete the message only after SQL succeeds.
      • Make the SQL operation idempotent, for example by recording the deletion identifier under a unique constraint.

    Service Bus provides the durable backlog that maximumWaitingRuns does not. If ordering is required only per business entity rather than globally, use a session-enabled queue and assign the entity key as the session ID. Logic Apps supports the sequential-convoy pattern for ordered session processing.^5^

    This architecture still requires load testing of the SharePoint ingestion side because a queue cannot recover an event that the SharePoint trigger never returned.

    Choosing maximumWaitingRuns

    For Standard single-tenant workflows:

    • Trigger concurrency range: 1–100.
    • Maximum waiting runs: up to 200.
    • The documented host defaults are 100 concurrent runs and 200 waiting runs.^3^^6^

    Therefore, maximumWaitingRuns: 100 is valid, but it is not intrinsically recommended for this workload. With runs: 1, increasing it to 200 only buffers another 100 workflow instances; it does not provide durable capacity for 5,000 events and can leave runs waiting for a long time.

    Use 100 only after confirming:

    • The worst-case processing time for 101 sequential runs is acceptable.
    • The trigger’s returned batch remains within concurrency-reduced debatching behavior.
    • Monitoring alerts cover waiting, failed, skipped, and throttled trigger runs.

    Also note that trigger concurrency cannot be disabled after it is enabled through the normal configuration path, so validate the workflow in a nonproduction environment first.^1^


    References

    1. Schema reference guide for trigger and action types in Azure Logic Apps
    2. SharePoint
    3. Limits and configuration reference for Azure Logic Apps
    4. Multi-region deployments for disaster recovery in Azure Logic Apps
    5. Send related Service Bus messages in a specific order by using a sequential convoy in Azure Logic Apps
    6. Edit app and host settings for Standard logic apps in single-tenant Azure Logic Apps
    AI-generated content may be incorrect. Read our transparency notes for more information.

    Was this answer helpful?

Your answer

Answers can be marked as 'Accepted' by the question author and 'Recommended' by moderators, which helps users know the answer solved the author's problem.