Gateway Issue

[cassidy]

Anyone out there having possible Gateway issues on the US West Coast?

Around 6 a.m. PST, my normal 2 minute Service to Gateway Dataflow ran for 2 hours without completing, can't get it to go through.

Updated/restarted Gateway.  Query runs fine locally without a Gateway/Service involved.

Thanks

[bcdobbs]

I had a weird issue yesterday where a normally quick data flowrepeatedly hung in the same way.

Mine was running in a premium per user workspace. I dropped the workspace to pro and then put it back and all was fine again. Did post on the forum but had no responses.

[cassidy]

This is a Premium Per Capacity workspace.  That's kinda terrifying to drop it to Pro and put it back as it's a critical workspace...any possible issues doing that?

[bcdobbs]

Mine was a small set of dataflows with no actual models! Afraid I couldn't comment on implications without knowing data volumes and any incremental load setups.

[bcdobbs]

Just to be clear the fix that worked for me might be completely unrelated. I wondered if my issue was something to do with the enhanced compute engine.

You could try turning it off on just that dataflow and back on?

As said, no firm envidence but it might help and is safer than changing something at workspace level.

[cassidy]

Tried both the Premium to Pro to Premium and the Engine with no luck.

What does work is rebuilding the Dataflow in a different Premium Workspace.

Datasets in the Workspace with issues seem to work, Dataflows do not.

[cassidy]

Okay, well it seems to now have started working.  I don't know why, all other Workspaces worked fine during this period, even those with Dataflows hitting the same source.  That leaves me thinking there was a 4-5 hour issue with this particular Workspace if that is even possible!

Weird morning

[cassidy]

It's back to going in the wronggg direction, each refresh is longer than the last.  Something is very strange.  I've opened a Ticket and will report back if I learn anything.

[cassidy · Accepted solution]

A mind-blowing 24 hours.

The issue resolved on it's own 24 hours later, Microsoft rep could not come up with a cause.

I had a copy of the Dataflow running in a different Workspace in the same Capacity, 0 issues over the 24 hours, while the original continued to perform terribly (as much as 5 hours longer to refresh).

This led us to believe it was literally the Workspace itself with an issue, yet nothing about that Workspace changed or differed from the others.  No idea.