New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.

Issue 632188 link

Starred by 2 users

Issue metadata

Status: Assigned
Owner:
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 2
Type: Feature

Blocking:
issue 667557


Participants' hotlists:
ops-backlog


Sign in to add a comment

Pipe machine version information through sysmon so we get metrics about our fleet

Project Member Reported by benhenry@chromium.org, Jul 27 2016

Issue description

In our meeting about updating the fleet, it was decided that the best next step was to get some data about where we are today.

https://docs.google.com/document/d/1irKwjhiZAOUJ8pLDF5zUSf9ArAIwRxBbJyj8dCsENLY/edit

Vadim offered to help pipe data through sysmon. We may need help on the graphing side.
 
Cc: dsansome@chromium.org chrishall@chromium.org
+ Chris and Dave, since it's very similar to the recent tsmon modification to send OS version info. Also feel free to grab this feature request if you are interested. It's low priority for me, will be implementing it slooowly.

-----

We can probably start with Linux, using once-an-hour collector in tsmon.

Here's my proposal for the stuff we want to collect.

Linux
-----
* Currently running kernel version ("cat /proc/version").
* Versions of following apt packages ("dpkg -l"). This is approximate list:
  ca-certificates
  chrome-remote-desktop
  curl
  facter
  git
  libc-bin
  libc6-*
  libstdc++6*
  linux-firmware
  linux-headers-generic
  linux-image-generic
  openjdk-*
  openssh-*
  openssl
  puppet
  python
  python-crypto
  ruby
  zlib*
Plus packages mentioned here: https://chromium.googlesource.com/chromium/src/+/master/build/install-build-deps.sh
* Time since last "apt-get update" call: "stat -c %y /var/lib/apt/periodic/update-success-stamp"
* Number of packages that can be upgraded (and number of security upgrades): "/usr/lib/update-notifier/apt-check"
* "Need reboot" flag.

OSX
-----
OSX version, including BuildVersion: "sw_vers"
Don't know what else... Have no idea how we update our Macs.


Windows
-------
* Kernel build number. 
* Number of pending updates?.. (Does it make sense in our environment?). Something like https://exchange.nagios.org/directory/Plugins/Operating-Systems/Windows-NRPE/Check-Windows-Updates-using-Powershell/details
* Perhaps identifiers of all recently installed updates (so we can asserts that an update such-and-such was installed on all our machines).
* "Need reboot" flag.
Labels: -Type-Bug Type-Feature

Comment 3 by benhenry@google.com, Oct 31 2016

Hey Vadim - when do you think you'll get to this?
Blocking: 667557
Thanks for the CC.

This is similar to my work in
https://bugs.chromium.org/p/chromium/issues/detail?id=623856
with the actual change happening in
https://codereview.chromium.org/2106953006/

My changes were mostly in infra/services/sysmon/system_metrics.py

As part of that work I also made some changes in
infra/services/sysmon/__main__.py
to only collect metrics every hour

So you should be able to re-use that logic by adding a call to your metric collection under
https://codesearch.chromium.org/chromium/infra/infra/services/sysmon/__main__.py?q=infra/services/sysmon/__main__.py&sq=package:chromium&dr&l=73
and then clearing your metric like I do, here
https://codesearch.chromium.org/chromium/infra/infra/services/sysmon/__main__.py?q=infra/services/sysmon/__main__.py&sq=package:chromium&dr&l=77

Naively I think the new proposed work here fits into this pattern.

Happy to help in anyway I can, don't think I have the capacity to take over this work at this moment.

Sign in to add a comment