> On May 23, 2017, 1:04 a.m., Joseph Wu wrote: > > What do you think about making fetch failures a sub-metric of > > `container_launch_errors`? The fetcher is only called during container > > launch. If fetching fails, container launch will also fail (and increment > > this counter). It might be useful to separate out fetch failures from > > other types of launch failures. This approach might be a bit more work to > > do so, since the parent metric lives inside the Agent, rather than the > > Fetcher process.
I like the idea of more specific launch errors, but that's not equivalent to what this metric gives. This lets you calculate the number of tasks that fetch assets as well as the proportion of asset fetches that fail. In the future I'd like to add more metrics under `containerizer/fetcher`; we could track disk usage, URI scheme, total URI fetch failures, cache hit and miss, etc. That's more complicated, so I'm proposing this to get something simple but useful. - James ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/59464/#review175740 ----------------------------------------------------------- On May 22, 2017, 9:15 p.m., James Peach wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/59464/ > ----------------------------------------------------------- > > (Updated May 22, 2017, 9:15 p.m.) > > > Review request for mesos and Mesos Reviewbot. > > > Bugs: MESOS-7524 > https://issues.apache.org/jira/browse/MESOS-7524 > > > Repository: mesos > > > Description > ------- > > Add the Fetcher metrics to track the number of fetch requests sent to > the Fetcher (`containerizer/fetcher/total`) and the number of errors > reported by the Fetcher (`containerizer/fetcher/errors`). > > > Diffs > ----- > > src/slave/containerizer/fetcher.hpp > 9e3018dc087ed55c61b2824d0105bc5339b83043 > src/slave/containerizer/fetcher.cpp > a910fea5a5556afb376524c5bb2ff98d7d84e611 > > > Diff: https://reviews.apache.org/r/59464/diff/2/ > > > Testing > ------- > > make check (Fedora 25) > > > Thanks, > > James Peach > >
