Closed nilzzzzzz closed 3 years ago
Thanks for reporting this. Could you please give us more information that could help replicating the issue in our end, like what version of Kong are you using and whether you are using any other plugins on these failing requests?
We are using Kong Version 2.1.4.
On the route which has most of the occurrences of the exception we have no other plugin than the global zipkin plugin.
I would like to share more information, but really no clue what correlates to this issue. I checked out the handler.lua and span.lua, but could not find a reason why a wrong timestamp could be passed to the annotate function..
I researched a bit more. This issue only occurs for users which are using a tool called addguard, which seems to use a proxy server for detecting malware / advertisements. No idea how that can influence the tracing in kong though..
we had the same issue, it looks like we had a typo or leftover in the span timestamp field which used only if the access phase is skipped
Hello @nilzzzzzz I think @Asafb26 's fix might be the source of your issue. I have merged it in master, it will come out on the next version of the plugin/Kong.
If you need to fix the issue right faster, I suggest you use the master branch of this plugin, which is stable.
@kikito thanks for the update and @Asafb26 thanks for the fix!
Hi,
i am having a weird issue in my kong production instance. A small percentage of requests fail with status code 400 (upstream server returned 200) and in those cases i see in the kong log an exception coming from here:
Any idea, why the annotation timestamp could be wrong or in which direction i should look?
Thanks in advance
Nils