Currently both transitlog-server and ui (express server) do very little logging. This makes troubleshooting quite difficult, specially during the deployment phase (but also during production)
Conventions used in other stacks (at least Transitdata):
output to System-out, which Docker swarm will redirect to correct place
For server side logging it would make sense to perhaps log all the requests with these information (if possible):
endpoint invoked (and since we're using graphql it could make sense to add some part of the payload?)
timestamp
HTTP result code
time elapsed to process the request?
possible other app level info if useful
for example more fields?
{
"http-result":200,
"processing-time":120 # ms
}
Also please add detailed logging specially in the beginning of the app (f.ex possible connection errors to db, etc) and also some message if server is shut down, etc?
Currently both transitlog-server and ui (express server) do very little logging. This makes troubleshooting quite difficult, specially during the deployment phase (but also during production)
Conventions used in other stacks (at least Transitdata):
example log output (you can add more fields but perhaps consider using the same naming conventions?)
For server side logging it would make sense to perhaps log all the requests with these information (if possible):
for example more fields?
Also please add detailed logging specially in the beginning of the app (f.ex possible connection errors to db, etc) and also some message if server is shut down, etc?