Currently there are 22k metrics like:
ipc_blobs_finality_voting_failure{blob_hash="227nstvgcjm4v6br4bicckixavemccu5d5fenev2oscakhmojkfa",subnet_id="/r314159/t410fntaew3dtef2zpfox2hlhy4cgrksdt5zjwg3hbfi"} 1
ipc_blobs_finality_voting_success{blob_hash="227nstvgcjm4v6br4bicckixavemccu5d5fenev2oscakhmojkfa",subnet_id="/r314159/t410fntaew3dtef2zpfox2hlhy4cgrksdt5zjwg3hbfi"} 2
The current approach results in high cardinality and leads to unnecessary high prometheus and grafana memory and costs. We can have just a single counter for the failure and log the failed hash.