Benchmarking tools #177
Benchmarking tools #177
Conversation
sergefdrv
force-pushed
the
bench
branch
2 times, most recently
from
1fcbe4b
to
44aafa3
Compare
|
@sergefdrv you'll need to rebase again if you want a green checkmark from the linter. |
I think it would make more sense once we have another atomic broadcast protocol to benchmakr. |
sergefdrv
force-pushed
the
bench
branch
2 times, most recently
from
b015ab5
to
28c3aef
Compare
|
@xosmig since I'm on vacation, could you please merge this PR if you're fine with how I address your comments? |
We will probably need to benchmark the availability layer soon (in fact, it could even make sense to do it now, with multisigcollector). |
Enjoy your vacation! :)
let's return to discussing this PR after you come back. |
|
Hi! I think this PR is a pretty good start for the benchmarking tooling. Of course, a lot of things can be extended and generalized. When we have some cycles to actually use it before Sergey comes back from vacation, we can see whether that's reason enough to merge it before generalization (or even generalize it ourselves), or continue improving it as a PR and only merge when more general. |
matejpavlovic
force-pushed
the
bench
branch
2 times, most recently
from
b718d4a
to
d06c96f
Compare
|
I will update the PR to use the new “SMR system” abstraction. |
Nodes gather the following statistics as they run: - the number of delivered requests, - the average number of delivered requests per second, - the average latency to deliver a request. Nodes periodically output statistic records in CSV format to the specified file or to the standard output.
Signed-off-by: Matej Pavlovic <matopavlovic@gmail.com>
Signed-off-by: Matej Pavlovic <matopavlovic@gmail.com>
This PR resolves #174.