When I was in school back home in India, my mom wouldn't buy a ping pong paddle for me to play with friends since it was expensive. So, I decided to cut down old notebook cardboard covers and I made one for myself by glueing 7-8 hard cardboard notebook covers.
I know, sounds lame but this is all I can remember. :D
I was expecting way too much in this article. But, I don't see anything about logging requestId's etc.
Sharing a few things briefly below I have learned over the past few years -
1. Log every single message for a transaction/request with the requestId to look at the entire picture. Ideally, that requestId would be used upstream/downstream too internally for finding exact point of failure in a sea of micro services.
2. Log all essential id's/PK's for that particular transaction. - makes debugging easier.
3. Log all the messages possibly in a json format, if the log aggregator you are using supports that. Parsing later for dashboards makes lives WAY EASIER and is more readable too. Might also reduce overall computations being run on the cloud to extract values you want -> hence, lesser cost.
4. Having error/warn logs is good, but having success/positive logs is equally important imo.
5. Oh, and be very very careful about what you end up logging. We don't want to log any sort of user PII. At one of my previous companies, there were mistakes made where we were logging user's phone numbers, addresses etc.
I am sure the community here already knows about these, and might have even more to add. Would love to hear what other people are doing.
I know, sounds lame but this is all I can remember. :D