Error in folsom metrics name
Hello,
Currently, the metrics name isn't the one in README, at least for folsom.
[[hackney_pool,erestcl_hackney_pool,take_rate],
[hackney_pool,erestcl_hackney_pool,queue_counter],
[hackney_pool,erestcl_hackney_pool,in_use_count],
[hackney_pool,erestcl_hackney_pool,free_count],
[hackney_pool,erestcl_hackney_pool,no_socket],
[hackney,total_requests],
[hackney,finished_requests],
[hackney,nb_requests]]
Regards.
hrm I don't follow, what should be the expected names? You mean in the metrics related to a pool?
bump
I'm having trouble with metric names as well. The problem I see here is the use of a list of atoms for metric names. When using {mod_metrics, folsom} in combination with folsomite the metric names are sent to a graphite host using a gen_tcp. Unfortunately, if the metric key is a list folsomite assumes the list is an iolist which fails for lists that contain atoms. I see this could be solved in folsomite but would require to iterate over all metric keys that are lists - including proper iolists. Interestingly, folsomite would handle this case properly if the metric key would be a tuple but I don't believe a tuple would work for exometer though.
I will convert all metrics name in string in the next version. It should fix any issue you see.
:+1:
I can't be sure this is what was being originally referred to, but from what I can gather I believe the issue is this states that the pool metrics are prefixed with hackney, but the actual metric is prefixed with hackney_pool, as can be seen here
This seems to be a case of confusion (at least it was for me), but I could also just be misreading the documentation. Happy to make a pr if this the metric names are indeed misdocumented