Not sure if the title is clear enough, but I'll try to explain in more detail.
If both timestamping and batch_window_size is enabled (respectively true and something greater than zero) I get an error on no case clause matching
14:18:36.504 [error] CRASH REPORT Process exometer_report_influxdb with 0 neighbours crashed with reason: no case clause matching {[erlang,memory],#{<<"host">> => "timmov-mbp.local"},#{total => 50843616},1459513116503950} in exometer_report_influxdb:maybe_send/5 line 261
This patch resets the timestamp so that for [:erlang, :memory] the timestamp is taken from the latest in the list (one of these: [:total, :processes]).
Not sure if the title is clear enough, but I'll try to explain in more detail.
If both timestamping and batch_window_size is enabled (respectively true and something greater than zero) I get an error on
no case clause matching
This happens when you define a metric as:
And a subscription as:
This patch resets the timestamp so that for
[:erlang, :memory]
the timestamp is taken from the latest in the list (one of these:[:total, :processes]
).The result of this patch is this (ngrep):
Which I think is mentioned in https://github.com/travelping/exometer_influxdb/issues/14#issuecomment-194902675
Please let me know if this is acceptable.