Unable to access browser's local storage. Some features such as undo or redo might not work until you clear browser's data

[joelzehring]

Recently, Power Automate has started to display a warning banner when editing a flow:

"Unable to access browser's local storage. Some features such as undo or redo might not work until you clear browser's data."

After clearing browser cache and other stored data, this warning still appears.

This coincides with an issue where Running a child flow fails because the instant trigger does not have access to location data.

Anyone else receive this warning?

[kiki_515]

Same here, this first time I encountered such error too 🙂 and anyone can help to advice  how to resolve it? Clearing browser cache, download & history done, but still same warning, I can't edit my flow now...haiz...

[BelindaBosman]

Yes, same here. Very annoying.  

[Anonymous]

This has recently started happening to me too. Clearing browser data doesn't fix it.

Also coincides with the issue when running a child flow. With experimental features turned off or on, the manual trigger appears to have all these metadata parameters (which are not explicitly established in the child flow):

In addition, with experimental features turned on, when I try to dynamically select a variable, the action collapses into an error:

FWIW, with experimental features turned off, it appears to at least let me select a variable dynamically.

EDIT: It's not consistent among flows. Some of them only mark a few of the 'extraneous' parameters as required:

[HEATFreight]

Same thing happening to me! Reset Edge cache with default setting and "all time". Did not fix the problem, but sure as heck was a nightmare to relaunch Edge with probably hundreds of tabs logged into various O365 junk.

I lost all my context! And every focused O365 tab in every window and some of the not-focused tabs as well ask for login creds. But once logged in you can use the browser's back button on extraneous login pages and refresh the content with newly cached credentials, however some tabs get stuck on a login page and the previous URL is lost! Unrelated I know, but really unfortunate side effect of being ADHD and following Microsoft's instructions that they link to in the error:


Unable to access browser's local storage. Some features such as undo or redo might not work until you clear browser's data.Learn more

 https://go.microsoft.com/fwlink/?linkid=2186936

[HEATFreight]

It seems like all my flows are affected. Even flows which are not Parent or Child flows are affected.

I'm even seeing this affecting flows both within Solutions and flows built outside of Solutions.

[jmarinuk]

I've also just started seeing this message this week.  Clearing the cache/browsing history has not helped.  I'm using Edge as my primary browser at work.

[JWiden297]

I'm also getting this message when trying to edit any Flow on my primary Edge profile. However, if I switch to a different profile or incognito tab I do not. Have tried clearing cookies/cache and restarting to no avail. 

[HEATFreight]

More strange behavior like what @Anonymous was seeing...

I have Parent/Child flows failing for broken input schemas, but my Child flow input schemas are all single text inputs (which contain the JSON string of the data I want to pass from Parent flow to Child flow). It's literally looking for exactly one input. But now –only on certain Child flows– I see these extra input fields coming out of nowhere:

I also saw a variation of the collapsed action error when I was trying to edit the formula for the input of a Select action. The mitigation is to directly paste the code using @{function()} notation instead of using the formula editor.

[larmccoy42]

I have seen the same error.  It appears to happen in Chrome, both with and without Ingnito mode.  I've also seen this error in Edge one all of my profiles.

[Anonymous]

No response from Microsoft? This seems to be happening suddenly and to a lot of people, and really messes with the development of flows. It'd be nice to get someone to at least officially acknowledge this...

[HEATFreight · Accepted solution]

Guys, due to this bug, you may find that some button-triggered flows suddenly have their intrinsic location properties as required inputs, causing the "Run a Child Flow" action to not run and the Parent flow to fail. For me, this has only happened with Child flows that I've just edited. Not every edited child flow behaves this way, but for me ONLY those I've edited since the bug appeared seem to be capable of exhibiting this behavior. The "Run a Child Flow" action in the Parent flow will look something like this:

The workaround is easy! For any affected flow, simply edit all Parent flows which run that particular Child flow, and put a 1 or something arbitrary in each of those unwanted required fields. You can make it work for the time being simply by giving each of them a value:

As far as I can see, the bug seems to be mostly superficial once you've dealt with this issue.

You may have trouble with the function editor, in which case you can just write out the function in a Notepad window using @{function()} notation, then paste the entire curly-bracketed code directly into the field.

I have also observed some weird stuff going on in Power Apps starting at the same time this bug began affecting things, but that may be entirely coincidental.

tl;dr: The two main impacts of the bug that I have observed both have easy workarounds, bolded above.

[Anonymous]

@HEATFreight, thanks for posting the workarounds. I also came to the same conclusion regarding the first workaround, but it's forcing me to litter my flows with all that crap. I suppose it's better than just waiting for MS to fix this, but they really should get on it...

[HEATFreight]

I noticed that the so-called "extraneous" parameters are deceptively required, despite the missing red asterisk!

I only have a couple affected Child flows across two environments. In my Dev environment, all those location fields show up as required (red asterisk). In my Prod environment, only the first field has a red asterisk. However, I get the same error for the Parent flow "Run a child flow" action in the Prod and Dev environments, unless I arbitrarily fill in ALL of those location fields, red asterisk or not.

I recall that environments have some kind of production vs. sandbox sort of setting. So it would seem that environments using the "production environment" setting are erroneously showing those fields as extraneous (on top of erroneously showing them at all).

Hope they fix this soon, but wonder if the workaround will break things after the fix? Hopefully whatever fix they deploy will deal with that.

[joelzehring]

In the absence of a solution from Microsoft, your post gets the nod as the solution to this issue. The same worked for us. Turning off Experimental Features also seems to have improved stability somewhat in the flow editor.

[Anonymous]

I feel like Microsoft still needs to weigh in on this.

@HEATFreight's workarounds are essentially just that...workarounds. We can't be developing flows by adding 1's to parameters that shouldn't be there. Keep in mind that if we continue doing so, and Microsoft eventually fixes the issue, a lot of those flows with the then-uneccessary 1's are going to start breaking no? And I don't even want to think about how that will affect any flows that have been sent over to production with those extraneous parameters.

[HEATFreight]

---

Keep in mind that if we continue doing so, and Microsoft eventually fixes the issue, a lot of those flows with the then-uneccessary 1's are going to start breaking no?

---

Yes, exactly this! Will they roll back the arbitrary data we used as inputs to fix the workaround? What could those possibly break? If you actually use those intrinsic location parameters for something, then inputting arbitrary data probably breaks your logic. Or even like in my case if you don't use the intrinsic location parameters for anything at all, something still might break after this issue has been resolved on the Power Automate side.

I have an ongoing support ticket and I keep asking them to reply to this thread with their suggestions, but I think that's crossing the streams. The Microsoft folks who would reply here are probably not the same as those who handle support tickets, most of whom are actually 3rd-party contractors not employed directly by Microsoft.

Here's the last communication I received in the email thread for my support ticket:

I hope you are having a good start of the week. 

My name is _____, I am one of the supervisors of the Power Apps and Power Automate team.

In this instance, I am contacting you, as I would like to schedule a call so we can review together the current state of your case, so we can reach a resolution as soon as possible and clarify any further questions.

What is your availability? My working hours go from 1 PM until 10 PM EST.

I will be waiting for your reply

Every time they ask me to get on yet another call with them, I reject their request and ask that they email or post the fix here. I've already been on a 1 to 2 hour screenshare with them where I showed a support engineer all the symptoms of the issue. Why can't the solution be emailed? Why would I have to schedule another call/screenshare? Honestly Microsoft's 3rd-party contractor support company has no idea what they are doing except for the most basic of issues.

And before that last email here are the troubleshooting steps they recommended:

I hope this email finds you well

I want you try the following troubleshooting steps, to see if this can mitigate the issue, please review the following document:

Troubleshooting startup issues for Power Apps - Power Apps | Microsoft Docs(section: Enable storage of third-party cookies and local data in your browser or app).
You have steps for multiple browsers / operative systems.

After you enabled the storage of third-party cookies and local data, please also follow the steps from the section "Configure Trust Zones for Internet Explorer" - Option 1. The settings defined on Internet Explorer affect multiple things in your local machine, even if you are not using Internet Explorer or Edge as your default browser.

This action plan have worked for other customers before with similar issues, please let me know regards your findings,

Our customers’ satisfaction is paramount to us. Please do not hesitate to contact me if you have further concerns or queries regarding this case.

Rest assured that we will provide you with the necessary support you deserve.

Here is the hyperlink from that message: Troubleshooting startup issues for Power Apps - Power Apps | Microsoft Docs

I gotta be honest, I have not really looked into this at all and have definitely not tried to follow their advice, because I think they are wrong and I do not believe their suggestion takes into account the novel nature of this issue. It seems they are supporting the issue as if it's a problem on my end, as if some setting or cookie issue or local data cache is the problem. Well, in my opinion we didn't all just coincidentally start having the same problems simultaneously. That doesn't totally eliminate the possibility that their troubleshooting steps could work, but it casts doubt as to whether they're approaching the issue from the right angle. I don't have time to be their guinea pig, especially when I've already mitigated the issues and returned to business as usual.

Maybe someone else wants to try their suggested fix and let us know how it goes, but I just do not have time to do it right now.

[HEATFreight]

Tata Consultancy Services Ltd is the garbage-tier 3rd-party support company who Microsoft contracts to sift through their support tickets.

The folks I've interacted with at Tata are always exceptionally kind and patient, and some of them are fairly knowledgeable, but they don't have the resources to deal with these sorts of highly nuanced and entirely novel issues.

It's probably not even their fault. Maybe it's merely the difference between 3rd-party support vs. real Microsoft support engineers. I've had the pleasure of dealing with a Microsoft support engineer and they are great.

And I think my screenshare was with an actual Microsoft support engineer, but now it seems they've passed the baton back to Tata.

Guess I'll eat my words if the troubleshooting steps that Tata suggested and which I've ignored thus far turn out to be the solution.

[Anonymous]

---

Well, in my opinion we didn't all just coincidentally start having the same problems simultaneously. That doesn't totally eliminate the possibility that their troubleshooting steps could work, but it casts doubt as to whether they're approaching the issue from the right angle. I don't have time to be their guinea pig, especially when I've already mitigated the issues and returned to business as usual.

---

Yeah, they need to look at this thread and understand that something randomly broke one day for multiple users and the fix is likely not on our end. And most of us paying customers already lose a lot of time with Power Automate's "quirks" to also spend time showing them what they need to debug on their own systems.

[cortese98]

Same issues people are reporting here. Anyone have an update?