The “updated knack record” trigger is broken in and of itself with the new performance update, it does not retrieve all information correctly from the record, I’ve yet to find to find a pattern regarding field types.
The two-step method that both of you describe works succesfully in my tests as well.
The fix I deployed is as you describe: to only retrieve the record ID and then have the second action step be the “Find record trigger” to obtain all record information.
This being said, it is clear that only our migrated applications are affected.
I am still in talks with the Zapier team and hope that the root cause of this change can be found and fixed in tandem with the Knack team.
Thanks @BradStevens - yes, adding a step to subsequently read the record in Zapier works. However, we have many subsequently steps that rely on the values and so it necessary to update all of these with the values from Step 2 (rather Step 1). Either way, it is a lot of work and I’ve started down the Make/Integromat path. As you mentioned, it is a bit of learning to move from Zapier to Make, but it works well. Testing is bit more clunky (no way that I can see to replay steps), and there are more steps to achieve the same outcome, but I think it has some advantages. I particularly like that you are no limited to 3 branches as you are are in Zapier.
@Pablo - I was able to recreate the same behaviour with a newly created app (which is already performance upgrade). I have the same problem with text concatenation fields.
I am having the same trouble with the knack-record-update where it is not retrieving any calculated field values, even tho I can see the values in the knack builder. This is causing havoc with my JavaScript that is looking for values in those fields. I’m also seeing incorrect summarization on fields that sum a connected record that have filters.
Yesterday we managed to get the “updated record trigger, then extract information from the found record action” up and going on Zapie yet…
The issue persists
Even extracting field values from the “found knack record” action does not prove to be a stable work-around. We’ve now encountered that in one of our applications even the values from an auto-increment field are shown as blank!
This “blank-field-value” issue is still arising, and we’ve noticed that it has changed behaviour unexpectedly throughout the day.
@KnackSupport,
It seems that the new performance upgrade has broken the way webhooks were working.
Derived fields like Auto-increment and Text Concatenate come across as blank from the webhook. Also connected fields like user name.
We were able to get the values by adding a 0.1 minute Delay between the “Updated Record in Knack” trigger and “Find Record in Knack” action.
You will also need to update all subsequent steps in the Zap to use the values retrieved in “Find Record in Knack”.
The “Find Record in Knack” action also has an annoying bug within Zapier. It will wrap around text which caused more problems. To fix we needed to also add the Text formatter in Zapier to strip out the HTML tags.
It all got a bit hard, and so we now rebuilding all of our Zaps using Make/Integromat.
Make/Integromat takes a bit of getting used to, but adds some powerful features which I think will work well for us in the long term.
@AdrianBell27622 some very experienced Make/Integromat users here so I encourage asking questions in this forum if needed, and the new Make community is very active and responds quickly in m experience.
Does Make use webhooks for it’s “search” and/or “find” methods?
Did your connections between Make - knack work without error these last few days?
@KnackSupport,
The work-around we used is failing intermittently in our zapier-knack connection, it seems that if we use an auto-increment field as an “internal object ID” the data is not able to be read.
We’re still porting our Zapier integrations over to Make, but the ones we have completed are (so far) working without issue. The updates also run significantly faster.
When you create a webhook in Make, you have the following options for knack event:
Record created
Record updated
Record deleted
This will just pass the Record ID, which you use then need to “Get a Record” and retrieve the Data.
When you create the Make Scenario, you have the following options for Knack:
Triggers:
As mentioned above by @AdrianBell27622
Make modules allow webhooks shown below.
In Zapier you can also fire a webhook on a form submission.
You can achieve the same result in Knack by sending a mailhook to Make.
This gives even more scope and flexibility as it can start a scenario on a form submit for a new record, or a record edit, even and action link.
Anywhere you can send an email in Knack you can start a Make scenario.
When you create a webhook in Make, you have the following options for knack event:
Our Knack - Zapier - Textmagic integration also stopped working with the upgrade August 9. We have 6 Zaps that run with a variety of filters. 4 of them continued to work fine and are still working fine. Those 4 looked for new records and then sent text messages to affected parties. The other 2 looked for updated records and previously sent text messages to affected parties. The updated records do not come across completely after the upgrade and therefor the filters don’t work and the zaps stop. I tried some work arounds and even rebuilt a test zap but nothing would come across.
I contacted both Knack and Zapier after I dug into the issue. Zapier replied quickly and let me know they had seen similar behavior. Knack took 5 business days (we are on a Pro Plan) to respond and let me know they were backed up due to issues from the upgrade and would let me know when they found something.
Obviously we are disappointed so I clicked around and found this entry. So today I carved out some time and created a couple Make Scenarios to replicate the Zaps that aren’t working. So far so good, and the tips on this thread helped immensely.
Thank you all for your insights and if I can provide any help, please drop me a line.
Its unbelievable that Knack is that irresponsible that they would destroy the integration with no concern at all. I have multiple apps using 30 plus Zaps tied to update a record. These are apps that I have over 1500 customers using them. I did see the work around but still unbelievable. Not professional at all.
This issue is affecting large swaths of your userbase causing irreparable damage in some applications / use cases.
The time, capital, and resources being spent by your customers including myself is beyond understanding considering the high fees some of us are subscribed to.
I’ve had no response from the @KnackSupport team since tuesday and have written almost every day requesting a status on the fix.
Do tell us IF this issue will be solved, and if so, WHEN. This way we can take the appropiate measures and not be left holding the bag.
As a just in general update to this issue, so everyone is on the same page, we had a fix that we were set to deploy last week for this and it needed further work, it’s the top known issue and should be complete soon, as soon as we deploy a fix we’ll make sure that the relevant page is updated, and if I’m on at the time I’ll update this thread as well.
We released a fix for the Zapier issues this afternoon. If you still run into problems with them please make sure to submit a ticket to support@knack.com but feel free to tag me here as well so we can get those tickets as fast as possible.
Apologies Tim,
we’ve got people on this one, but any extra data that you can provide in an email to support@knack.com (unless you have, if you have can you give me the ticket number or DM me your e-mail so I can search for it) and I will make sure it gets added to the current ticket to help get it resolved asap. The information we’re looking for is:
Zapier setup screenshots - As thorough as you can provide
Result of the Zap
View that is used in the Zap trigger or action - a link to this would be great
History of the execution of the Zap (This can be retrieved from Zapier.)
Any errors you see
Thank you for pinging to bring it to my attention, I’m dipping into having a more active role in the forums and I did not even see the notification that there was a ping. Sorry that it caused a delay in getting back to you for more of this information.
Hi Adrian, we have had the same issue since the performance upgrade. We had an app that has worked for years and now is unusable. Its frustrating to say the least when you have put so much time in and your whole bussiness is running on it.
