Hi there!
Your platform is a truly good IPSAAS, that helps us connect different systems together. But its logging features are really a big stopping force, that just makes overall problem searching difficult.
My request would be towards extending the current logging system to be far more useful than it is currently, instead of having not-so useful data (like process id, trace_id, console_command etc.): Have the exact error trace in every error log message, and if possible, the data state at that exact moment (more than enough, if the failing data is kept).
An example:
Lets say, that I have an incoming configuration, named “Product sync”. It has multiple entity transformers (3) inside, with lots of setters, conditionals, mappers etc inside of them. (namely Transformer1, Transformer2, Transformer3)
Lets say, that a scheduled run of “Product sync” fails due an error. If for some reason, in one of the setters a non-existing path is referenced, an error log will be created, which will say probably something like “Key xyz was not found in data”, and that the error has happened in the “Product sync” incoming.
It will not always mention which exact entity transformer has caused the issue, and even if it does, it DEFINITELY will not mention; which EXACT data transformer (aka. which setter, for example) has caused the issue. Or in other cases, it doesn’t always elaborate on how exactly is that a problem, just that it happened - like presented by this picture:
From the above picture, I got pretty much no information about what has happened, where it has happened, just that it was an error, and that it can be tied to one specific incoming config - of which actually contained 6 transformers, with a long list of data transformers each.
If we would have an exact error trace that shows us the exact data transformers that failed, or at least an indication about how the data looked like then; it would be god-sent.
Please keep in mind, that I’m not talking about the overall debug features - this is strictly about the logging system (so explaining error messages on runs, that are e.g. scheduled using cron - the logs are the only data from which we can debug).
Thank you for your attention, and congrats on building this platform! Keep up the good work.
Best regards,
Gergő
7digits - backend developer