Closed aboodman closed 3 years ago
@phritz it seems like it should not be hard to get this to run on either an dedicated instance in ec2 or at least isolated? I would like to add this to near term work. wdyt?
Definitely seems important to have reliable metrics for the things that matter most, so yes, let's do it.
Something we should address: that we dont get fine-grain commit results for repc changes. We should definitely fix this so we don't have to go digging into releases to find when a thing happened. Maybe a smoke perf test on the repc side to guide the search or test-releasing every commit in repc so we can run the benchmarkes on the js side.
moved populate degredation issue, which is separate, to https://github.com/rocicorp/replicache-sdk-js/issues/218
Closing for now.
We are running the tests on our own EC2 instance which should be more stable.
This is close to being useful:
https://replicache-sdk-js-perf-git-perf-data.rocicorp.vercel.app/
... but not useful. See for example the dip in populate perf around 63c44cc24cdeaa40053b533625f3bf0b380ed985. Is it real? If so which commit? Hard to tell.
@phritz it seems like it should not be hard to get this to run on either an dedicated instance in ec2 or at least isolated? I would like to add this to near term work. wdyt?