Excessive Execution Time for For Each and Conditional Actions in Azure Standard Logic App

Question

Excessive Execution Time for For Each and Conditional Actions in Azure Standard Logic App

Shabnam Shaikh 20

Hi Team,

We are facing a performance issue in our Azure Standard Logic App workflows where built-in actions such as For Each, Condition, and sometimes Do Until are taking an unusually long time to execute.

Issue Details:

Even when processing only 1 or 2 records, the workflow execution takes around 10–15 minutes.
One of our workflows recently took 78 minutes to process only 10 records.
Under normal conditions, the same workflow usually completes within 3-4 minutes for 10–20 records.

Observations:

The issue is mainly observed with built-in control actions such as:
- For Each
- Condition
- Do Until
The actions inside these control blocks execute within normal expected durations. However, the overall execution time of the parent control action is significantly higher.
Example:
- Actions inside a For Each loop may complete within a few seconds collectively.
- But the For Each action itself shows a total execution duration of 10–15 minutes or more.

Pravallika KV 16,700 Reputation points Microsoft External Staff Moderator

2026-05-21T13:19:58.17+00:00
Hi @Shabnam Shaikh ,

It sounds like your Standard Logic App is spending most of its time in the “wrapper” of control actions (For Each / Condition / Do Until) rather than in the actions themselves. Normally you’d expect a 10-20-item loop to finish in a few minutes, so 15-78 minutes suggests there’s some host-level or configuration overhead going on.

Here’s what you can try:

Check your loop concurrency

By default, For Each runs iterations in parallel (Concurrency = Off). If your degree of parallelism is set to 1, the engine may be waiting behind internal throttles or cold starts for each iteration.

To increase parallelism: open your For Each in the designer → Settings → Concurrency control → turn it On → bump Degree of parallelism up (for example, to 10 or 20).

If you actually want sequential processing, explicitly set the loop to Sequential so that the engine only waits once up front instead of re-negotiating per iteration.

Look for cold starts or host warm-up delays

In a Standard (single-tenant) Logic App, each workflow runs on an underlying Functions host. If your host scales down between runs, the first few iterations can pay a cold-start penalty.

Consider using deployment slots or keeping your host “warm” via a simple ping-type run every few minutes.

Split large or nested loops into child workflows

If you have nested For Each or Do Until blocks, each level adds orchestration overhead. Splitting a big workflow into smaller parent/child workflows can reduce the time the parent control action spends orchestrating.

Monitor and diagnose with built-in metrics

Enable Application Insights for your Standard Logic App (host.json) so you can see per-action telemetry, cold-start counts, and concurrency metrics.

Review the run history in the portal and click into each control action to see breakdowns of “wait,” “evaluation,” and “execution” times.

Verify you’re not hitting any service limits or throttling

The default limits for a single-tenant Logic App are up to 100 K array items and up to 50 concurrent loop iterations. If you’ve configured extremes outside defaults or share the same API connection across many apps, you could see throttling.

If throttling is suspected, create separate connections or slow down the loop concurrency.
Deepak Tambe 0 Reputation points

2026-05-21T13:52:56.4833333+00:00
@Pravallika KV

Thank you for the detailed recommendations

We would like to highlight that everything was working as expected earlier, and this issue has only started occurring over the past two days. No changes have been made from our end in terms of workflow design, loop concurrency settings, infrastructure, configuration, or deployment.

The specific issue we are observing is that built-in control actions such as For Each, Condition, and sometimes Do Until are taking an unusually long time to execute, even when the actual actions inside those controls complete quickly.

For example:

Small loops with limited iterations are experiencing significant delays.

The majority of the workflow runtime is being spent within the orchestration/control actions themselves rather than connector calls or business logic execution.

This behavior was not present previously and started suddenly without any application-side modifications.

Since the issue began recently and without any changes from our side, could you please help verify whether there are:

Any ongoing platform-side performance degradations,

Recent infrastructure or runtime updates,

Or known issues affecting Standard Logic Apps orchestration performance?

Please let us know if you need any additional run history details, correlation IDs, or timestamps from our side for further investigation.

Regards,
Deepak
Simarjot Singh 0 Reputation points

2026-05-25T01:19:18.78+00:00

Hey Shabnam, we experienced this recently starting 19th May in Australia East region. We followed the fix mentioned here which is binding AzureFunctionsJobHost_extensionBundle_Version env variable to [1.161.25, 1.161.25]. This is not a permanent fix but has worked for us so far. https://learn.microsoft.com/en-us/answers/questions/5895298/timeouts-and-excessive-execution-times-on-workflow

2 answers

Your answer

Deepak Tambe 0 Reputation points

2026-05-21T13:52:56.4833333+00:00

@Pravallika KV

Thank you for the detailed recommendations

We would like to highlight that everything was working as expected earlier, and this issue has only started occurring over the past two days. No changes have been made from our end in terms of workflow design, loop concurrency settings, infrastructure, configuration, or deployment.

The specific issue we are observing is that built-in control actions such as For Each, Condition, and sometimes Do Until are taking an unusually long time to execute, even when the actual actions inside those controls complete quickly.

For example:

Small loops with limited iterations are experiencing significant delays.

The majority of the workflow runtime is being spent within the orchestration/control actions themselves rather than connector calls or business logic execution.

This behavior was not present previously and started suddenly without any application-side modifications.

Since the issue began recently and without any changes from our side, could you please help verify whether there are:

Any ongoing platform-side performance degradations,

Recent infrastructure or runtime updates,

Or known issues affecting Standard Logic Apps orchestration performance?

Please let us know if you need any additional run history details, correlation IDs, or timestamps from our side for further investigation.

Regards,
Deepak
Simarjot Singh 0 Reputation points

2026-05-25T01:19:18.78+00:00

Hey Shabnam, we experienced this recently starting 19th May in Australia East region. We followed the fix mentioned here which is binding AzureFunctionsJobHost_extensionBundle_Version env variable to [1.161.25, 1.161.25]. This is not a permanent fix but has worked for us so far. https://learn.microsoft.com/en-us/answers/questions/5895298/timeouts-and-excessive-execution-times-on-workflow

Answer 1

Pravallika KV 16,700 Microsoft External Staff Moderator

Hi @Deepak Tambe / @Shabnam Shaikh ,

The hotfix was deployed to all the public regions, and you can now safely move back to the default extension bundle version - [1.*, 2.0.0). You should be able to get the hotfix through simple restart if you are still affected.

Hope this helps!

If the resolution was helpful, kindly take a moment to click on User's image and click on Yes for was this answer helpful. And, if you have any further query do let us know.

Pravallika KV 16,700 Reputation points Microsoft External Staff Moderator

2026-05-27T09:52:43.1666667+00:00

@Shabnam Shaikh ,Following up to see if the provided answer was helpful. If this answers your query, do click Accept Answer =>Yes, and upvote it. If you have any further queries do let us know.

Answer 2

Hi Shabnam Shaikh

Thanks for reaching out on Microsoft Q&A!

The pattern your experience could be a Logic App Standard issue which isn't the execution time, but orchestration overhead within the workflow runtime. There could be several causes you could check:

Storage latency/throtteling: under the hood, Logic App Standard relies on Azure Storage (tables, queues and blobs) and random spikes could affect built-in actions. Please check the metrics if you find throttling errors, latency or availability issues. When it's a General Purpose v2, you could move to Premium or enable zone-redundant storage (ZRS). Also, when applicable, try to avoid sharing storage with heavy apps (perhaps other workflows running on the same storage?
App Service Plan resourcing: when you have a small SKU (for example B1, S1, etc.) you could experience resource exhaustion. Please check whether you can find spikes of CPU, memory or Thread Count, and if so, scale up to a higher plan. Think of it like your own computer, when it's busy it simply hangs.
When your Logic App isn't used frequently it could experience a "cold start" due to recyling of workers. You could try to enable 'Always On'.
Are large payloads used? This could "eat up" the resourcing of your Logic App, where basically the same applies as the point noted above, only this one isn't as transparent as you would expect. Everything's loaded in memory, so when working with large data sets it can grow above limits very fast.

Please click ‘Accept answer’ if you think my answer is helpful. Feel free to drop additional queries in the comments below!

Kind regards,

Sonny

Share via

Excessive Execution Time for For Each and Conditional Actions in Azure Standard Logic App

2 answers

Your answer