Skip to content
Projects
Groups
Snippets
Help
Loading...
Help
Support
Keyboard shortcuts
?
Submit feedback
Contribute to GitLab
Sign in / Register
Toggle navigation
slapos
Project overview
Project overview
Details
Activity
Releases
Repository
Repository
Files
Commits
Branches
Tags
Contributors
Graph
Compare
Issues
0
Issues
0
List
Boards
Labels
Milestones
Merge Requests
0
Merge Requests
0
Analytics
Analytics
Repository
Value Stream
Wiki
Wiki
Snippets
Snippets
Members
Members
Collapse sidebar
Close sidebar
Activity
Graph
Create a new issue
Commits
Issue Boards
Open sidebar
Cédric Le Ninivin
slapos
Commits
eeab64e0
Commit
eeab64e0
authored
May 27, 2016
by
Alain Takoudjou
Browse files
Options
Browse Files
Download
Email Patches
Plain Diff
monitor stack collect: initialise empty files if collector db doesn't exist
parent
3afd3d95
Changes
1
Hide whitespace changes
Inline
Side-by-side
Showing
1 changed file
with
47 additions
and
25 deletions
+47
-25
stack/monitor/scripts/collect.py
stack/monitor/scripts/collect.py
+47
-25
No files found.
stack/monitor/scripts/collect.py
View file @
eeab64e0
...
...
@@ -262,51 +262,73 @@ def appendToJsonFile(file_path, content, stepback=2):
jfile
.
seek
(
position
)
jfile
.
write
(
'%s}'
%
',"{}"]'
.
format
(
content
))
def
initProcessDataFile
(
file_path
):
with
open
(
process_file
,
'w'
)
as
fprocess
:
data_dict
=
{
"date"
:
time
.
time
(),
"data"
:
[
"date, total process, CPU percent, CPU time, CPU threads"
]
}
fprocess
.
write
(
json
.
dumps
(
data_dict
))
def
initMemoryDataFile
(
file_path
):
with
open
(
mem_file
,
'w'
)
as
fmem
:
data_dict
=
{
"date"
:
time
.
time
(),
"data"
:
[
"date, memory used percent, memory used"
]
}
fmem
.
write
(
json
.
dumps
(
data_dict
))
def
initIODataFile
(
file_path
):
with
open
(
io_file
,
'w'
)
as
fio
:
data_dict
=
{
"date"
:
time
.
time
(),
"data"
:
[
"date, io rw counter, io cycles counter, disk used"
]
}
fio
.
write
(
json
.
dumps
(
data_dict
))
if
__name__
==
"__main__"
:
parser
=
parseArguments
()
if
not
os
.
path
.
exists
(
parser
.
output_folder
)
and
os
.
path
.
isdir
(
parser
.
output_folder
):
raise
Exception
(
"Invalid ouput folder: %s"
%
parser
.
output_folder
)
collector
=
RessourceCollect
(
parser
.
collector_db
)
date_scope
=
datetime
.
now
().
strftime
(
'%Y-%m-%d'
)
stat_info
=
os
.
stat
(
parser
.
output_folder
)
partition_user
=
pwd
.
getpwuid
(
stat_info
.
st_uid
)[
0
]
# Consumption global status
process_file
=
os
.
path
.
join
(
parser
.
output_folder
,
'monitor_resource_process.data.json'
)
mem_file
=
os
.
path
.
join
(
parser
.
output_folder
,
'monitor_resource_memory.data.json'
)
io_file
=
os
.
path
.
join
(
parser
.
output_folder
,
'monitor_resource_io.data.json'
)
resource_file
=
os
.
path
.
join
(
parser
.
output_folder
,
'monitor_process_resource.status.json'
)
status_file
=
os
.
path
.
join
(
parser
.
output_folder
,
'monitor_resource.status.json'
)
if
not
os
.
path
.
exists
(
parser
.
collector_db
):
print
"Collector database not found..."
initProcessDataFile
(
process_file
)
initMemoryDataFile
(
mem_file
)
initIODataFile
(
io_file
)
with
open
(
status_file
,
"w"
)
as
status_file
:
status_file
.
write
(
'{"cpu_time": 0, "cpu_percent": 0, "memory_rss": 0, "memory_percent": 0, "io_rw_counter": 0, "date": "", "total_process": 0, "disk_used": 0, "io_cycles_counter": 0, "cpu_num_threads": 0}'
)
with
open
(
resource_file
,
"w"
)
as
resource_file
:
resource_file
.
write
(
'[]'
)
exit
(
1
)
collector
=
RessourceCollect
(
parser
.
collector_db
)
date_scope
=
datetime
.
now
().
strftime
(
'%Y-%m-%d'
)
stat_info
=
os
.
stat
(
parser
.
output_folder
)
partition_user
=
pwd
.
getpwuid
(
stat_info
.
st_uid
)[
0
]
process_result
,
memory_result
,
io_result
=
collector
.
getPartitionComsumptionStatus
(
partition_user
)
resource_file
=
os
.
path
.
join
(
parser
.
output_folder
,
'monitor_process_resource.status.json'
)
label_list
=
[
'date'
,
'total_process'
,
'cpu_percent'
,
'cpu_time'
,
'cpu_num_threads'
,
'memory_percent'
,
'memory_rss'
,
'io_rw_counter'
,
'io_cycles_counter'
,
'disk_used'
]
resource_status_dict
=
{}
if
not
os
.
path
.
exists
(
process_file
):
with
open
(
process_file
,
'w'
)
as
fprocess
:
data_dict
=
{
"date"
:
time
.
time
(),
"data"
:
[
"date, total process, CPU percent, CPU time, CPU threads"
]
}
fprocess
.
write
(
json
.
dumps
(
data_dict
))
initProcessDataFile
(
process_file
)
if
not
os
.
path
.
exists
(
mem_file
):
with
open
(
mem_file
,
'w'
)
as
fmem
:
data_dict
=
{
"date"
:
time
.
time
(),
"data"
:
[
"date, memory used percent, memory used"
]
}
fmem
.
write
(
json
.
dumps
(
data_dict
))
initMemoryDataFile
(
mem_file
)
if
not
os
.
path
.
exists
(
io_file
):
with
open
(
io_file
,
'w'
)
as
fio
:
data_dict
=
{
"date"
:
time
.
time
(),
"data"
:
[
"date, io rw counter, io cycles counter, disk used"
]
}
fio
.
write
(
json
.
dumps
(
data_dict
))
initIODataFile
(
io_file
)
if
process_result
and
process_result
[
'total_process'
]
!=
0.0
:
appendToJsonFile
(
process_file
,
", "
.
join
(
...
...
@@ -324,7 +346,7 @@ if __name__ == "__main__":
)
resource_status_dict
.
update
(
io_result
)
with
open
(
os
.
path
.
join
(
parser
.
output_folder
,
'monitor_resource.status.json'
)
,
'w'
)
as
fp
:
with
open
(
status_file
,
'w'
)
as
fp
:
fp
.
write
(
json
.
dumps
(
resource_status_dict
))
# Consumption Ressource
...
...
Write
Preview
Markdown
is supported
0%
Try again
or
attach a new file
Attach a file
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Cancel
Please
register
or
sign in
to comment