My scenario worked well for a few days then all of a sudden it just stopped working last Friday 12th July.
I am sending a query to Monday.com through a http post and then with the returned data I am sending it to a Lamda function through a http post to do some logic then sending the returned .csv file to a SharePoint and emailing it to my customer in the form of a report.
The first http post works fine and returns the data as expected but then when I send it to AWS Lambda I get a timeout, the reason I think it’s got something to do with make is that I replicated the http post in postman, and it works fine.
The URL has been triple checked, all of the advanced options are default.
I also have another scenario in another account that uses a different AWS account, and that scenario has the same error and it started happening on the same day.
I have already contacted support, but it’s been 3 working days and I have had no response so I am hoping someone here might be experiencing or has experienced the same issue.
Yes, very weird, there were no auth restrictions at all, it’s a public endpoint on one of the AWS accounts, was going to add authentication later when I got some time.
I did hear back from support yesterday and i was asked to try the timeout rule in the advanced settings, There does seem to be some conflict in your documentation re the default timeout,
and even if this was the issue the scenario returns the timeout after 12 seconds,
It’s possible this has never come up before because most people would just use the Amazon Lambda module and create the connection (is what I have decided to do), I was just so used to using http to do everything I didn’t even think to use a premade module.
Would still be good for someone to try and replicate this issue and confirm if it’s on the Make.com side or the AWS side.
And to answer your question, yes, the scenarios for both make.com accounts in both AWS accounts all started to fail on the same day, I couldn’t pinpoint the exact time, but it was early afternoon when I started getting all the failure emails. It was working for 3 days then just stopped working through no change to either application.
Thanks for all these details.
If it works with the AWS Lambda Make App, it would be interesting to compare what it sends for the request with what you do with the HTTP App.
Have you already used the Integromat Dev Tool in Chrome? Besides some tools, it can show details of requests and responses that occur in the scenario when you test-run it.
And I think that the timeout you mentioned is only for connections that worked but with the server being too long to reply. In your case, it seems that the call doesn’t even manage to get any http connection.
If you can try the Integromat Dev Tool, can you let me know the outcome?
Before I made the switch I did find timeout settings on the AWS side in the configuration menu in the function, I boosted them both but it still didn’t work for me, however I don’t see that being the issue since it works with the LAMDA module with the same (AWS) timeout and memory settings.
@Ben_Channon can you please also try the Lambda Make App and also leave the HTTP module just after and test with the Integromat dev tool? The goal would be to try and figure out if there’s any difference in the url or method or params, and investigate why it started to fail.
In the meantime I will check if there’s anything new in Make.
I’m not sure what “Lambda” is sorry. Is assume it’s a software @Neville is using.
My issue is with HTTP module same as Neville, but i’m calling an API we build custom for another software. As with Neville, everything worked perfectly fine until the 10th July then we got the same error.
Postman runs the Exactly same Request successfully without error
Another MAKE account on US1 runs the Exactly same Request successfully without error
Integromat DEV Tool has the same error with my account, shows no extra info, just that same ETIMEDOUT error.
If it works from the us1 instance, it seems to be an issue with the outbound ip of the eu1/2 zones, but since it works for @Neville with the Lambda App, I wanted to check all possible root causes.
We’ve had to resort to this follow rubbish workaround:
Sending all HTTP API requests to ZAPIER via their module,
Running the API call Request from the Zapier HTTP module.
Sending the ENTIRE response back to MAKE and processing the rest of the scenario from there.
I cannot stress how wildly inefficient this is, not to mention the doubling of potential error points
@Benjamin_from_Make if you can help in anyway or need anything please reach out and ask.
For us this means our entire make profile is dead in the water without a resolution as everything else we have in MAKE depends on these API calls as the initial source of data.
I wanted to flag this, because there are several other users who are stuck because of this error. Might be nothing, but it doesn’t look like a coincidence. It seems like there is an error on Make’s side, as one of these users was able to do the exact same request outside of Make. I have asked what region their Make account is in:
I got around to installing and using the Devtool extension, The Lambda function uses the AWS ARN (amazon resource name) because it’s connected to the AWS account. But the HTTP needs to use the URL.
I’m all out of ideas, unfortunately I’m no developer, I use Make.com because it’s sold as “low code” and I understand the irony of saying that when I’m sending data to lambda functions to run nodejs… I just hope now that there is some additional evidence this is deeper than just my issue (Thanks @Henk-Operative ) someone with some real expertise (from Make) can now take over and maybe replicate the issue in a sandbox account and get to the bottom of the issue.
Starting on the 12th of July AEST, all calls using the HTTP request module to our Student Management System have started failing with the exact same error message.
Our student management system (destination to the API calls) is also an AWS instance.
Identical calls with the same parameters are successful and fully operational when executed via Postman/Zapier/Hubspot, as well as Make’s HTTP Request module on the US1 server.
My ticket has been open since the 12th, which is when we first detected the errors (#1074553).
It’s likely that this issue is impacting a significant number of users who rely on the HTTP module for custom integrations with unsupported applications.
If a solution is not immediately available, could we at least receive an acknowledgment of this issue on Make Status.
Thanks for your feedbacks. As stated yesterday, the issue is currently under investigation by experts, trying to figure out why calls are blocked for the European zones of Make.
There’s clearly something blocking calls in specific cases, and as I’m sure you all know, these types of issues are hard to investigate. But they are taken seriously and worked on.
Thank you @Neville for your tests, I’m sure it will help Make team with their investigation.
just fyi, I tested with Postman, from France, and my calls to Lambda via HTTP are blocked, it won’t work from my computer unless I use a VPN.
This confirms that the issue is not coming from Make, but AWS blocking some regions.
Support is still investigating what could be happening.
In the meantime, could you please check again AWS Lambda to see if they introduced a new parameter that limits regions that can call their Lambda endpoints?
