You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Continuation of this discussion on the critical leak that affected 2.8.7 to 2.12.1.
We need a way to detect if something like this happens again.
It would be a good idea IMO to have general benchmarks in any case, so we can detect how our changes affect loopback's throughput. We could also fail the benchmark outright if a leak event is caught by node-memwatch.
@raymondfeng What do you want to test in the benchmark? I'm thinking we should make loopback-benchmark public and get some community input on this. Opinions?
As discussed with @raymondfeng and @ritch, it has been decided that we will add benchmarks to various perspectives of LoopBack (module/unit level, integration level, etc). The purpose of this issue has now changed to a spike on how to move forward with regards to tooling and statistics gathering. The findings will be captured in a wiki.
Activity
bajtos commentedon Mar 30, 2015
Relabelled per our discussion with @raymondfeng, @ritch and @superkhau
raymondfeng commentedon Mar 30, 2015
@superkhau Would you like to create a simple benchmark test that leverages https://github.com/lloyd/node-memwatch to see if it helps?
superkhau commentedon Mar 30, 2015
Sure, adding to TOB for planning.
superkhau commentedon Apr 8, 2015
@raymondfeng What do you want to test in the benchmark? I'm thinking we should make loopback-benchmark public and get some community input on this. Opinions?
superkhau commentedon Apr 8, 2015
As discussed with @raymondfeng and @ritch, it has been decided that we will add benchmarks to various perspectives of LoopBack (module/unit level, integration level, etc). The purpose of this issue has now changed to a spike on how to move forward with regards to tooling and statistics gathering. The findings will be captured in a wiki.
superkhau commentedon Apr 8, 2015
Closing this in favour of #1291