According to those guidelines, metrics "...should have a suffix describing the unit, in plural form. Note that an accumulating count has total as a suffix, in addition to the unit if applicable."
4 of the metrics:
dex_block_io_read_bytes
dex_block_io_write_bytes
dex_network_rx_bytes
dex_network_tx_bytes
... are counters and so I believe they should be suffixed to:
dex_block_io_read_bytes_total
dex_block_io_write_bytes_total
dex_network_rx_bytes_total
dex_network_tx_bytes_total
Furthermore, there is this metric dex_memory_total_bytes which is declared as a counter. This is part of the container spec and it can't change for the lifetime of the container. But if you recreate the container you can use a different limit, which if set to a lower value would increase the old value by that amount. In that sense, I believe we should change it to a gauge. If we don't change to a gauge, we could at least rename it to dex_memory_total_bytes_total, for the same reasons as above.
Let me know if you are interested in a PR to take care of this.
Hi there. There are guidelines on metric and label names on Prometheus documentation: https://prometheus.io/docs/practices/naming.
According to those guidelines, metrics "...should have a suffix describing the unit, in plural form. Note that an accumulating count has
totalas a suffix, in addition to the unit if applicable."4 of the metrics:
... are counters and so I believe they should be suffixed to:
Furthermore, there is this metric
dex_memory_total_byteswhich is declared as acounter. This is part of the container spec and it can't change for the lifetime of the container. But if you recreate the container you can use a different limit, which if set to a lower value would increase the old value by that amount. In that sense, I believe we should change it to a gauge. If we don't change to a gauge, we could at least rename it todex_memory_total_bytes_total, for the same reasons as above.Let me know if you are interested in a PR to take care of this.