Activities in pipeline timing out

[Ber]

Hi all

We have a pipeline, where some of the activities in some runs times out without any apparent reason. It has been an intermittent problm for some time, but the last few days, it happens every time we run the pipeline.

The step we first saw timing out, is a web activity calling an auth url through the gateway. Thinking it could be an inrrastructure problem, i disabled that step, but then a lookup activity, reading from a lakehouse times out instead.

The problem does not seem to be related to throttling.

Any input to what could be happening here is greatly appreciated?

[Ber]

@Ugk161610 Thank you for your suggestions, that helped.

[Ugk161610 · Accepted solution]

Hi @Ber ,

We’ve seen this same behavior a few times in our Fabric pipelines — especially with Web and Lookup activities timing out randomly, even when the setup and load are fine.

From our experience, this usually happens due to temporary backend or gateway session instability in Fabric. The activity connection sometimes drops before completion, which triggers a timeout even though the environment itself is healthy.

Here’s what helped for us:

• Increased the timeout duration for the affected activities.

• Added a retry policy (2–3 retries with a small delay).

• Checked the gateway health and ensured it wasn’t restarting or reconnecting.

• Adjusted the schedule time slightly to avoid peak load hours.

After applying these changes, the timeouts became much less frequent. Microsoft has already acknowledged similar cases, and improvements are being rolled out to make session handling more reliable.

Hope this helps!

Gopi Krishna