Impl Logger AddContext summand #162
Conversation
these images will now only show after this branch is merged
hs-abci-extra/src/Network/ABCI/Server/Middleware/ResponseLogger.hs
Outdated
Show resolved
Hide resolved
hs-abci-extra/src/Network/ABCI/Server/Middleware/RequestLogger.hs
Outdated
Show resolved
Hide resolved
hs-abci-examples/nameservice/src/Nameservice/Modules/Nameservice/Keeper.hs
Outdated
Show resolved
Hide resolved
| @@ -1,23 +1,147 @@ | |||
| # nameservice | |||
|
|
|||
| ## Metrics | |||
| ## Metrics via Prometheus | |||
There was a problem hiding this comment.
Before commenting on any of this, is there any way to have kibana pull metrics from prometheus in the same way datadog does. This would be good to fix some confusing asymmetry between server metrics and application metrics, and would also be easier to explain that you can now use either Kibana or datadog to view anything you want, or both.
There was a problem hiding this comment.
Good question. I’ll investigate.
There was a problem hiding this comment.
@blinky3713
It seems MetricBeat is an option from the ELK folks that can scrape prometheus exporters and servers. Given we're already using ELK stack, this seems to be a pretty good option.
Similar to the datadog agent, it also has a docker-compose.
There was a problem hiding this comment.
There is a problem with checkTx / deliverTx with how this is now. Namely, if checkTx is actually running transactions without committing state, then it's going to log the same event to ES twice. We need to set a context for check versus deliver, and probably need to set context for the transaction hash as well so things are easily searchable.
I think it may make more sense to leave this unsolved here and instead fix it in #167 since transactions are further isolated in that PR
Closes #161.