0xFugue
|
c069923bd0
|
Data collection: docker stats and host-proc (#106)
* sysfs, docker stats, docker ps
* .gitignore
* updated file names
* cpu, mem, i/o
* .gitignore
* .gitignore
* added scheduler
* pruned mem/blk measurement
* cpu, mem, blk, net stats
* single node containers one
* renaming,bug fix, signal handling
* code clean up; added all the necessary xlations to the logs
* cpustat, pid2docker
* remove temp. files
* tear_down, permission exception
* updating .gitignore
* param validation, sudo blk reads
* added output dir
* comments
* .gitignore & rename
* .gitignore & rename
* de-escalate privileges
* proc monitor
* container id to host end of virtual interface
* complete code rewrite with minimal sudo section
* set default iface for non-docker wakus
* added signal handlers and clean shutdown
* added 3 different per-process network stats
* +VmData, +VmStk, handling of veth pairs, optimal cpu stats
* +VmHWM, net1=net3, net2
* +VmHWM, net1=net3, net2
* pre merge
* pre merge
* .gitignore
* adjusted paths & stats only nim/go waku
* monitor.sh
* adding docker wait, nim-waku:nwaku-trace2
* TTY bug
* copy config/metric dirs to logs
* cleanup.sh
* procfs monitor updates
* run.sh cleanup
* run.sh metrics_infra cadvisor
* cleanup fix
* empty output dir
* host-proc
* added cadvisor conditionals
* enable WLS wait
* signal WLS once the measuring infra is up
* spin_up wls_cid
* remove print/exit
* renaming
* graceful signal exit, reduce logging freq
* stat only running waku instances
* comments, remove prints
* de-duplicate docker calls
* deduplicate docker calls in procfs.py
* multi nodes
* adding Jordi's signal'ng code and timing
* minimise the setup time
* split and move sudo ahead
* run.sh/procfs/docker stats named pipe synchornisation
* jack-up the max# open filehandles
* 100k file handles
* log timings, set resource limits
* complete rewrite of pid2did
* added start time + waku versions
* requirements.txt
* assert, stat collection duration
* main.sh -> dstats.sh
|
2023-04-20 00:16:02 +05:30 |