My Custom Integration’s sync apparently failed, but there is no information about WHY:
I am consistently seeing this sync error.
Please help me understand how to debug this issue.
Here is a different kind of sync error – still no explanation:
Error generating Report:
Is it really too much to ask, as a paying customer, for someone from Fibery team to look into why your backend is throwing this error, and help me understand what is going on?
I am very skeptical that my integration app is actually passing invalid JSON to Fibery; but if it is, it would be helpful to have a little more context to understand WHEN this is happening.
What is “state STRING1”?
What record was being processed when this error occurred?
What was the last successfully processed record?
What are the limits on data returned from the integration app? The last packet sent had a 1.4MB payload.
Is there a limit in the number of entities in a DB, or total DB size?
Waiting for dev person to be back and help
(most likely tomorrow)
If we missed smth in community - please, don’t hesitate to text in the intercom chat, there we are definitely faster
Sorry for delay. I checked log for details. Looks like your payload is too large. Please implement and use paging in your custom connector. Fibery Integration Apps API
- What are the relevant limits I am exceeding?
- Is there any place I can see these error messages? e.g. “Payload Too Large”
To be honest it is a first time we met the issue with the size of data. We will need some time to found the part of where the limitation applied and make the error clear for users.
Sorry for inconvenience, we will notify about results.
I am all about pushing the limits LOL
@Oleg, I have reorganized my Fibery Integration app to limit the max records that I send to Fibery in a single response, to prevent the packet size from getting too large.
But I sometimes still see this error:
Below you can see some info about the sync data packets that my Integration app sent to Fibery which resulted in the above error (Note “PDT” = UTC-7) – the largest packet was only 262K (and previously I have seen Fibery integration successfully ingest 1.4MB packets):
It appears something else is happening here, besides just “too-large response packet”?
Thanks for letting us know. We will investigate the issue. Could you please provide the code of your connector? Probably it may help our developers to understand the problem.
Is it possible that 3rd party service replied with 500 error code and HTML response and you resend it as is to our service but with HTML status code 200?
What I see from our logs:
- response from connector is received
- library was not able to parse it into JSON
- error is risen (
Invalid JSON (Invalid UTF-8 character at position 0 in state STRING1)
here is more details
<html><head> <meta http-equiv="content-type" content="text/html;charset=utf-8"> <title>500 Server Error</title> </head> <body text=#000000 bgcolor=#ffffff> <h1>Error: Server Error</h1> <h2>The server encountered an error and could not complete your request.<p>Please try again in 30 seconds.</h2> <h2></h2> </body></html> . Invalid JSON (Invalid UTF-8 character at position 0 in state STRING1)
So, I can assume, that connector sent error as 200 response - result is failed sync. For this case I suggest to send it as error and include in error body flag
tryLater: true or send with 429 status code.
Thank you for looking into this issue. What you suggest makes the most sense, however my connector code already checks for errors received from Zoho (data source), and this error checking does work – I have seen it log errors that are retuned from Zoho, handled correctly.
In this case the connector detected no errors from Zoho. Also I looked at the captured data packets that I am returning to Fibery and I do not see anything wrong there (all good JSON).
Yes it is possible that I am passing this HTML to Fibery in a data response, but I do not see how it could be happening.
Are you able to see any more context about this HTML on your side – ?
- what was the request sent to the data connector (URL)
- exact time/size of request/response
I don’t track full response but I track invalid chunk.
Could u please check where the following symbol is located? Symbol - � (UTF \uFFFD)
Thanks @Evgeni_Kisel –
- do you have a timestamp associated with receiving that chunk?
- do you know which DB was being synced?
- Are any Unicode chars a problem for Fibery if they exist in the data records retuned from the Integration connector?
Currently I am only filtering out “unprintable” Unicode chars, using the following regex (Note that
\uFFFD is not in the list):
- type name is
Jun 9, 2023 @ 20:09:25.065(time in GMT+3)
- not sure. I don’t know where this unicode char is used in your response. I’ve tested as a part of name and it works ok (see screenshot below it’s an integration where I add that symbol at the end of each name)
- The Unicode chars are present in text/string field values.
- I filtered out
\uFFFDand sync was better - but there were still sync errors.
- I filtered out
\uFF28and now this seems much better.
My sync logs (logged data sent from my connector app to Fibery) have NO weird Unicode characters in them, but I still get this error when syncing the “Contacts” DB.
Please check if you can find any clues on your side – TIA
Played around by trying different responses. I was able to get this error only when my response just contain that single symbol.
Just for verification. Is it possible that connector just send response (empty response) with this UTF-8 symbol?
could you also log your responses? for example you can group them by header