Details
-
Improvement
-
Resolution: Unresolved
-
Major
-
master
-
None
Description
ns_server is going to pass us a port (--stats-port) where we should expose Prometheus metrics.
Name | Type | Labels | Description |
cont_bk_buckets_monitored | gauge | Number of buckets currently being monitored due to continuous backup being on | |
cont_bk_vbuckets_monitored | gauge | bucket | Number of vBuckets monitored |
cont_bk_gaps_filled | counter | bucket vbid |
Number of times we have filled a gap between two vBucket UUIDs |
cont_bk_backed_up | counter | bucket vbid vbuuid |
Number of bytes that have been uploaded |
cont_bk_upload_time | histogram | bucket vbid vbuuid |
Time spent uploading for given vBucket UUID |
cont_bk_compacted_files_uploaded | counter | bucket vbid vbuuid |
How many compacted files have been uploaded. If this is >1 then we cannot guarantee that a restore would allow to any point in time |
cont_bk_disk_overhead | guage | bucket vbid vbuuid |
Number of bytes that are being kept on disk, that would otherwise be removed, because of continuous backup. i.e. the byte size of hard linked SSTables that have been removed from the normal data directory |
cont_bk_holes | counter | bucket vbid vbuuid reason |
Incremented when a hole (i.e. missing data has been found) |
Attachments
Issue Links
- blocks
-
MB-63224 [ContBk] The user should be alerted when there is a hole in the backup
- Open