-
Notifications
You must be signed in to change notification settings - Fork 1.6k
[Feature] Query Log #4944
Copy link
Copy link
Closed
Labels
area/operationsRelated to operational aspects of the DB, including signals, flags, env vars, etc.Related to operational aspects of the DB, including signals, flags, env vars, etc.kind/featureSomething completely new we should consider.Something completely new we should consider.status/acceptedWe accept to investigate/work on it.We accept to investigate/work on it.
Metadata
Metadata
Assignees
Labels
area/operationsRelated to operational aspects of the DB, including signals, flags, env vars, etc.Related to operational aspects of the DB, including signals, flags, env vars, etc.kind/featureSomething completely new we should consider.Something completely new we should consider.status/acceptedWe accept to investigate/work on it.We accept to investigate/work on it.
Experience Report
What you wanted to do
Analyze if my query generated any problems in my cluster Or if third devs have abused with too complex queries without analyzing the capacity of the cluster. Without having to use Jaeger Tracing.
What you actually did
Hard to debug without Jaeger or clues (like logs).
Why that wasn't great, with examples
Sometimes (Is not always) when I make queries too complex, with many features and wide parameters. My cluster crash. A clue via logs can help start debugging.
I added it to the backlog, if it is not feasible, please close this issue.
BTW, it would be interesting for this logging to be separated from the main Dgraph binary log. Something like Prometheus(?) or a file. We add a flag that appends query logs to a file.