Provide a way to disable metrics logs in simulation
- Dominant language
- C++
- Stars
- 16.7k
- Forks
- 1.6k
- Avg merge
- 1d 20h
- Merged PRs (30d)
- 126
Description
When debugging timeouts in simulation, the trace log file contains a lot of repeated metrics loggings, like `ProcessMetrics` or `MachineMetrics`, that in most cases would lower the signal to noise ratio a lot. By providing a way to disable metrics logs in simulation, it's easier to locate the phase where the code got stuck by reading the trace files.
Maybe we can simply add a new server knob that could be set via the simulation start command that would basically disable the `systemMonitor()` call done in `workerServer()` if the program is in simulated mode.
Contributor guide
Research direction
Start at the workerServer() path and inspect the systemMonitor() call when the program is in simulated mode. Trace how simulation start-command server knobs are defined and passed through. Done means a simulation option suppresses ProcessMetrics and MachineMetrics trace logging without changing normal-mode monitoring; verify with the relevant simulation trace output.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- cpp
- Domain
- distributed-systems, observability
- Issue type
- Feature
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 45/100