Add client statistics #75
Conversation
|
I feel pretty good about the types but definitely want the names looked at closely. |
| {"filestotal", "inodes_maximum", "The maximum number of inodes (objects) the filesystem can hold", s.gaugeMetric}, | ||
| {"kbytesavail", "kilobytes_available", "Number of kilobytes readily available in the pool", s.gaugeMetric}, | ||
| {"kbytesfree", "kilobytes_free", "Number of kilobytes allocated to the pool", s.gaugeMetric}, | ||
| {"kbytestotal", "kilobytes_capacity", "Capacity of the pool in kilobytes", s.gaugeMetric}, |
There was a problem hiding this comment.
I've got myself rethinking these metrics now (all the kilobytes_* ones). Should they actually be *_kilobytes instead?
There was a problem hiding this comment.
That does make sense, especially from a consistency standpoint. Echoing an offline discussion, we want to save this for another PR, correct?
sources/procfs.go
Outdated
| {"kbytesfree", "kilobytes_free", "Number of kilobytes allocated to the pool", s.gaugeMetric}, | ||
| {"kbytestotal", "kilobytes_capacity", "Capacity of the pool in kilobytes", s.gaugeMetric}, | ||
| {"lazystatfs", "lazystatfs", "Returns '1' if lazystatfs (a non-blocking alternative to statfs) is enabled for the client", s.gaugeMetric}, | ||
| {"max_easize", "max_easize", "Maximum Extended Attribute (EA) size in bytes", s.gaugeMetric}, |
sources/procfs.go
Outdated
| {"stats", "write_minimum_size_bytes", writeMaximumHelp, s.gaugeMetric}, | ||
| {"stats", "write_maximum_size_bytes", writeMaximumHelp, s.gaugeMetric}, | ||
| {"stats", "write_bytes_total", writeTotalHelp, s.counterMetric}, | ||
| {"xattr_cache", "xattr_cache", "Returns '1' if extended attribute cache is enabled", s.gaugeMetric}, |
There was a problem hiding this comment.
maybe rename xattr_cache_enabled?
sources/procfs.go
Outdated
| {"max_read_ahead_mb", "max_read_ahead_megabytes", "Maximum number of megabytes to read ahead", s.gaugeMetric}, | ||
| {"max_read_ahead_per_file_mb", "max_read_ahead_per_file_megabytes", "Maximum number of megabytes per file to read ahead", s.gaugeMetric}, | ||
| {"max_read_ahead_whole_mb", "max_read_ahead_whole_megabytes", "Maximum file size in megabytes for a file to be read in its entirety", s.gaugeMetric}, | ||
| {"statahead_agl", "statahead_agl", "Returns '1' if the Asynchronous Glimpse Lock (AGL) for statahead is enabled", s.gaugeMetric}, |
There was a problem hiding this comment.
statahead_agl_enabled? Or maybe even worth expanding out to statahead_asynchronous_glimpse_lock_enabled...though that is super long...
roclark
force-pushed
the
wip-client-jobstats
branch
from
6dee465 to
5e88e5b
Compare
|
Added |
sources/procfs.go
Outdated
| {"kbytesfree", "kilobytes_free", "Number of kilobytes allocated to the pool", s.gaugeMetric}, | ||
| {"kbytestotal", "kilobytes_capacity", "Capacity of the pool in kilobytes", s.gaugeMetric}, | ||
| {"lazystatfs", "lazystatfs_enabled", "Returns '1' if lazystatfs (a non-blocking alternative to statfs) is enabled for the client", s.gaugeMetric}, | ||
| {"max_easize", "max_easize_bytes", "Maximum Extended Attribute (EA) size in bytes", s.gaugeMetric}, |
There was a problem hiding this comment.
Worth adding an underscore to make it max_ea_size_bytes?
roclark
force-pushed
the
wip-client-jobstats
branch
2 times, most recently
from
1b11fa6 to
6a03511
Compare
sources/procfs.go
Outdated
| {"max_read_ahead_per_file_mb", "max_read_ahead_per_file_megabytes", "Maximum number of megabytes per file to read ahead", s.gaugeMetric}, | ||
| {"max_read_ahead_whole_mb", "max_read_ahead_whole_megabytes", "Maximum file size in megabytes for a file to be read in its entirety", s.gaugeMetric}, | ||
| {"statahead_agl", "statahead_agl_enabled", "Returns '1' if the Asynchronous Glimpse Lock (AGL) for statahead is enabled", s.gaugeMetric}, | ||
| {"statahead_max", "statahead_max", "Maximum window size for statahead", s.gaugeMetric}, |
There was a problem hiding this comment.
Should we make this statahead_maximum to comply with the precedent we just started with our other _maximum statistic?
sources/procfs.go
Outdated
| {"max_easize", "max_ea_size_bytes", "Maximum Extended Attribute (EA) size in bytes", s.gaugeMetric}, | ||
| {"max_read_ahead_mb", "max_read_ahead_megabytes", "Maximum number of megabytes to read ahead", s.gaugeMetric}, | ||
| {"max_read_ahead_per_file_mb", "max_read_ahead_per_file_megabytes", "Maximum number of megabytes per file to read ahead", s.gaugeMetric}, | ||
| {"max_read_ahead_whole_mb", "max_read_ahead_whole_megabytes", "Maximum file size in megabytes for a file to be read in its entirety", s.gaugeMetric}, |
There was a problem hiding this comment.
Should these three be maximum_*?
There was a problem hiding this comment.
I'm all for it, but do you mean the four max_ metrics? If it's just two, which ones and for what reason just for my curiosity?
There was a problem hiding this comment.
eh, yeah, I meant four. I edited the comment to say three, but clearly I can't count. Let's rephrase more clearly:
We should change any case of max to maximum instead. :)
There was a problem hiding this comment.
Haha, all good! Just wanted to clarify!
Signed-Off-By: Robert Clark <robert.d.clark@hpe.com>
roclark
force-pushed
the
wip-client-jobstats
branch
from
6a03511 to
0d2619e
Compare
Phase 1 of fixing #69.
Signed-Off-By: Robert Clark robert.d.clark@hpe.com