5 ms·

I'm surprised there's no mention of tagging with some kind of correlation ID. I find that even generating a quick one-off GUID at the beginning of the request

I'm surprised there's no mention of tagging with some kind of correlation ID. I find that even generating a quick one-off GUID at the beginning of the request (to be used across the entire request) makes tracing server logs substantially easier. I just tag every log (and exception) with that ID, so when searching with grep or Splunk or whatever, I have a quick and easy way to disambiguate which log is connected to other logs.

wainstead 5y ago
And there's a standard for it: https://www.w3.org/TR/trace-context/ https://www.w3.org/TR/trace-context/ Many libraries and tools support the standard now.
- tombert 5y ago
  And now I feel embarrassed for not knowing about this...Thanks for the link!
  - wainstead 5y ago
    Don't, I only learned about it recently, and the standard itself is quite young!
kerblang 5y ago
Just an opinion, but I tend to think of this as distributed monolithic development... the more log correlation is needed, the more likely a requirement has been chopped up and then duct-taped back together over HTTP.
- NovemberWhiskey 5y ago
  Not sure I follow that logic? If you have a heavily used service, it's likely to have many concurrent requests logging into the same log stream and you absolutely need a request correlation identifier to be able to thread the log statements for a single request back together. That's essential even if you don't actually call any other services as part of your implementation.
  - hinkley 5y ago
    CorrelationIDs aren’t just good for hard failures. They’re also good for figuring out why your backend started getting 20% more traffic on Monday for no reason. Someone introduced a duplicate call somewhere. But where?
  - kerblang 5y ago
    Late reply: This scenario is reasonable and I usually handle it by logging the thread ID in multi-threaded systems, which amounts to a "free" global variable.
- jniedrauer 5y ago
  What about distributed systems? I had to reach for this tool recently to trace requests as they propagate through up to 5 separate devices. Not everything is a web app.

wainstead 5y ago

And there's a standard for it: https://www.w3.org/TR/trace-context/ https://www.w3.org/TR/trace-context/ Many libraries and tools support the standard now.

tombert 5y ago
And now I feel embarrassed for not knowing about this...Thanks for the link!
- wainstead 5y ago
  Don't, I only learned about it recently, and the standard itself is quite young!

kerblang 5y ago

Just an opinion, but I tend to think of this as distributed monolithic development... the more log correlation is needed, the more likely a requirement has been chopped up and then duct-taped back together over HTTP.

NovemberWhiskey 5y ago
Not sure I follow that logic? If you have a heavily used service, it's likely to have many concurrent requests logging into the same log stream and you absolutely need a request correlation identifier to be able to thread the log statements for a single request back together. That's essential even if you don't actually call any other services as part of your implementation.
- hinkley 5y ago
  CorrelationIDs aren’t just good for hard failures. They’re also good for figuring out why your backend started getting 20% more traffic on Monday for no reason. Someone introduced a duplicate call somewhere. But where?
- kerblang 5y ago
  Late reply: This scenario is reasonable and I usually handle it by logging the thread ID in multi-threaded systems, which amounts to a "free" global variable.
jniedrauer 5y ago
What about distributed systems? I had to reach for this tool recently to trace requests as they propagate through up to 5 separate devices. Not everything is a web app.