Error 429 in Power Automate: Understanding and Fixing Throttling

Error 429 (Too Many Requests) in Power Automate means that a flow has exceeded the request limit of a connector or the platform within a short time window and is therefore being throttled. The most common cause is a connector with a tight time limit, such as the SharePoint connector with an upper limit of 600 actions per minute, or a flow with many parallel runs or loop iterations. The solution usually lies in reducing concurrency control, filtering data more strictly, and replacing recurring individual actions with batch operations. As of: July 2026.

What causes error 429 in Power Automate?

Error 429 occurs when a flow sends more requests to a connector within a time window than its service protection mechanism allows. According to the official documentation on Platform limits and throttling the service then returns error code 429 with a message such as "Rate limit exceeded. Try again in 27 seconds". This connector throttling is independent of the daily action limits of your license plan: a flow can therefore be throttled even though the daily quota is far from exhausted. Typical triggers are loops with many iterations, several flow instances running simultaneously, or a single connector being used in parallel by multiple flows.

Which limits should you keep in mind?

The exact figures vary depending on the connector and license, but some reference values are clearly documented. The SharePoint connector, for example, limits the number of actions to 600 per minute, per connection across all flows and not per individual flow, as described in the Power Automate Licensing FAQ describes. Additionally, a platform-wide action burst limit of currently 100,000 actions per flow within five minutes applies. Independent of this, there are daily API request limits that depend on your license plan and apply to all flow runs of a user within 24 hours. Important for troubleshooting: both successful and failed actions count toward these limits, as do retries and additional requests caused by pagination.

How do you fix 429 errors through concurrency control?

The most direct lever is to reduce the degree of concurrency in the trigger or in loops. Microsoft recommends exactly this step in the guide to "Too Many Requests" in Power Automate for a 429 error: in the trigger or action settings under "Concurrency Control", the degree of concurrency can be limited. By default, concurrency control is disabled for triggers, but it can be set to a value between 1 and 100, with 25 concurrent runs as the default value when enabled. For "Apply to each" loops, a separate concurrency value applies, which according to the limits for flows defaults to 1 and can be increased up to 50, or deliberately kept low to avoid errors. If a flow repeatedly hits the connector limit, deliberately lowering these values is often the fastest remedy, even though the flow then runs somewhat slower.

How does batching help avoid throttling permanently?

Batching reduces the number of individual requests that can be throttled in the first place. Instead of executing a separate action for each record in a loop, according to the guide on Avoiding Anti-Patterns several create or update operations can be bundled into a single HTTP request via batch requests, provided the respective connector supports this. For services without batch support, controlled parallel processing in the "Apply to each" loop is the alternative, according to Microsoft with up to 50 records simultaneously. It is also worth looking at the filter query and top count parameters of many connectors in order to retrieve fewer records from the outset, as well as at nested loops, which can quickly multiply the number of requests. Anyone who wants to make their Power Automate environment fundamentally more stable will find Power Automate consulting from NordFlux support for flow architecture, license selection, and troubleshooting recurring throttling.

Frequently asked questions about Power Automate throttling (429)

Why do I get 429 errors even though my daily action limit has not been reached?

Because connector throttling works independently of the daily API request limits. Every connector, such as SharePoint with its limit of 600 actions per minute, enforces its own short-term time window, which you can exceed even with low daily usage if many actions are executed in a short time.

Does a more expensive license plan help against 429 errors?

A higher license plan primarily increases the daily action quota, not the connector-specific throttling limits. Since the 429 error is usually triggered by connector throttling, adjusting concurrency and flow design is more effective than a license upgrade in most cases.

What is the difference between the action burst limit and connector throttling?

The action burst limit is a platform-wide upper limit of currently 100,000 actions per flow within five minutes, while connector throttling applies per individual connector service and varies greatly depending on the service. Both can independently lead to a 429 error, which is why the cause is best narrowed down using the exact error message and the flow run history.

Can I simply bypass 429 errors with automatic retries?

Automatic retries do not solve the underlying problem, because failed actions and retries also count toward the request limits. It makes more sense to address the cause through concurrency control, filtering, and batching, rather than simply letting the flow retry more often.