|
|
DescriptionUpdate perfbot sheriffing docs.
Include some information about chrome infra monitoring, and about milo.
Committed: https://crrev.com/052761c8caa9b11702a4548eb83b6483c1f3e084
Cr-Commit-Position: refs/heads/master@{#421960}
Patch Set 1 #
Total comments: 4
Patch Set 2 : Rebase and comments #Messages
Total messages: 19 (7 generated)
martiniss@chromium.org changed reviewers: + benhenry@chromium.org, sergeyberezin@chromium.org, sullivan@chromium.org
PTAL +sergeyberezin for the link to the internal monitoring dashboard. Is this allowed?
lgtm
LGTM + a couple of nits. Thanks! https://codereview.chromium.org/2380913002/diff/1/tools/perf/docs/perf_bot_sh... File tools/perf/docs/perf_bot_sheriffing.md (right): https://codereview.chromium.org/2380913002/diff/1/tools/perf/docs/perf_bot_sh... tools/perf/docs/perf_bot_sheriffing.md:85: [go/chrome-infra-mon](http://go/chrome-infra-mon). To debug offline buildslaves, nit: this is the first time I see that go/ link :-) We have too many for the same page (go/chrome-infra-viceroy, go/chrome-infra-graphs, ...). Let's standardize on vi/chrome_infra. https://codereview.chromium.org/2380913002/diff/1/tools/perf/docs/perf_bot_sh... tools/perf/docs/perf_bot_sheriffing.md:86: you can look at the "Individual machine" dashboard, under the "Machines" nit: provide direct link, e.g. vi/chrome_infra/Machines/per_machine . In fact, this link can go instead of the top-level link above (the top-level page is not very helpful for this paragraph). Another potentially interesting page is vi/chrome_infra/Buildbot/per_pool . Searching for 'perf' gives a list of relevant slave pools like this one: http://vi/chrome_infra/Buildbot/per_pool?pool=master.chromium.perf%3ALinux+Bu... The top graph shows the number of configured and connected slaves + the load on the pool.
estaab@chromium.org changed reviewers: + estaab@chromium.org
lgtm for milo section
https://codereview.chromium.org/2380913002/diff/1/tools/perf/docs/perf_bot_sh... File tools/perf/docs/perf_bot_sheriffing.md (right): https://codereview.chromium.org/2380913002/diff/1/tools/perf/docs/perf_bot_sh... tools/perf/docs/perf_bot_sheriffing.md:85: [go/chrome-infra-mon](http://go/chrome-infra-mon). To debug offline buildslaves, On 2016/09/29 at 18:23:01, Sergey Berezin wrote: > nit: this is the first time I see that go/ link :-) We have too many for the same page (go/chrome-infra-viceroy, go/chrome-infra-graphs, ...). > > Let's standardize on vi/chrome_infra. Done.
The CQ bit was checked by martiniss@chromium.org
The patchset sent to the CQ was uploaded after l-g-t-m from sergeyberezin@chromium.org, estaab@chromium.org, sullivan@chromium.org Link to the patchset: https://codereview.chromium.org/2380913002/#ps20001 (title: "Rebase and comments")
CQ is trying da patch. Follow status at https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...
(forgot to include one comment) https://codereview.chromium.org/2380913002/diff/1/tools/perf/docs/perf_bot_sh... File tools/perf/docs/perf_bot_sheriffing.md (right): https://codereview.chromium.org/2380913002/diff/1/tools/perf/docs/perf_bot_sh... tools/perf/docs/perf_bot_sheriffing.md:86: you can look at the "Individual machine" dashboard, under the "Machines" On 2016/09/29 at 18:23:01, Sergey Berezin wrote: > nit: provide direct link, e.g. vi/chrome_infra/Machines/per_machine . In fact, this link can go instead of the top-level link above (the top-level page is not very helpful for this paragraph). Done > > Another potentially interesting page is vi/chrome_infra/Buildbot/per_pool . Searching for 'perf' gives a list of relevant slave pools like this one: > > http://vi/chrome_infra/Buildbot/per_pool?pool=master.chromium.perf%3ALinux+Bu... > > The top graph shows the number of configured and connected slaves + the load on the pool. I don't think this is very useful for perf sheriffs. sullivan@, do you think it would be? My impression was that there isn't a big pool of machines that perf needs to monitor, which is what that page is good for. I didn't include a link to that.=
On 2016/09/29 20:08:35, martiniss wrote: > (forgot to include one comment) > > https://codereview.chromium.org/2380913002/diff/1/tools/perf/docs/perf_bot_sh... > File tools/perf/docs/perf_bot_sheriffing.md (right): > > https://codereview.chromium.org/2380913002/diff/1/tools/perf/docs/perf_bot_sh... > tools/perf/docs/perf_bot_sheriffing.md:86: you can look at the "Individual > machine" dashboard, under the "Machines" > On 2016/09/29 at 18:23:01, Sergey Berezin wrote: > > nit: provide direct link, e.g. vi/chrome_infra/Machines/per_machine . In fact, > this link can go instead of the top-level link above (the top-level page is not > very helpful for this paragraph). > Done > > > > > Another potentially interesting page is vi/chrome_infra/Buildbot/per_pool . > Searching for 'perf' gives a list of relevant slave pools like this one: > > > > > http://vi/chrome_infra/Buildbot/per_pool?pool=master.chromium.perf%3ALinux+Bu... > > > > The top graph shows the number of configured and connected slaves + the load > on the pool. > > I don't think this is very useful for perf sheriffs. sullivan@, do you think it > would be? My impression was that there isn't a big pool of machines that perf > needs to monitor, which is what that page is good for. I didn't include a link > to that.= Yeah, we don't really monitor how the load is distributed among a pool of machines, just whether things are breaking on a single machine.
> > http://vi/chrome_infra/Buildbot/per_pool?pool=master.chromium.perf%3ALinux+Bu... > > > > > > The top graph shows the number of configured and connected slaves + the load > > on the pool. > > > > I don't think this is very useful for perf sheriffs. sullivan@, do you think it > > would be? My impression was that there isn't a big pool of machines that perf > > needs to monitor, which is what that page is good for. I didn't include a link > > to that.= > > Yeah, we don't really monitor how the load is distributed among a pool of machines, just whether things are breaking on a single machine. I was suggesting pool graphs so you can see how many machines are actually online and connected to the master vs. how many are expected to. If the numbers don't match, this would be an indication that some machines are broken in a pool. The other per_machine console only shows if a specific machine is online (not even the slave process - just the OS), and you can't see the overall health of a fleet or a pool at a glance. But it's up to you.
The CQ bit was unchecked by commit-bot@chromium.org
Failed to apply the patch. On branch working_branch Your branch is up-to-date with 'origin/refs/pending/heads/master'. nothing to commit, working tree clean
The CQ bit was checked by martiniss@chromium.org
Message was sent while issue was closed.
Description was changed from ========== Update perfbot sheriffing docs. Include some information about chrome infra monitoring, and about milo. ========== to ========== Update perfbot sheriffing docs. Include some information about chrome infra monitoring, and about milo. Committed: https://crrev.com/052761c8caa9b11702a4548eb83b6483c1f3e084 Cr-Commit-Position: refs/heads/master@{#421960} ==========
Message was sent while issue was closed.
CQ is trying da patch. Follow status at https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...
Message was sent while issue was closed.
Patchset 2 (id:??) landed as https://crrev.com/052761c8caa9b11702a4548eb83b6483c1f3e084 Cr-Commit-Position: refs/heads/master@{#421960} |