[jira] [Commented] (HBASE-11062) htop
[ https://issues.apache.org/jira/browse/HBASE-11062?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16848258#comment-16848258 ] Toshihiro Suzuki commented on HBASE-11062: -- +1 for hbtop. [~elserj] {quote} You have a patch/PR for what you demo'ed (even with the LGPL dependency)? {quote} Not yet. I need to do some refactoring of the code and add some tests. Maybe, I can create a PR for it next week or in 2 weeks. > htop > > > Key: HBASE-11062 > URL: https://issues.apache.org/jira/browse/HBASE-11062 > Project: HBase > Issue Type: New Feature > Components: hbase-operator-tools >Reporter: Andrew Purtell >Assignee: Toshihiro Suzuki >Priority: Major > > A top-like monitor could be useful for testing, debugging, operations of > clusters of moderate size, and possibly for diagnosing issues in large > clusters. > Consider a curses interface like the one presented by atop > (http://www.atoptool.nl/images/screenshots/genericw.png) - with aggregate > metrics collected over a monitoring interval in the upper portion of the > pane, and a listing of discrete measurements sorted and filtered by various > criteria in the bottom part of the pane. One might imagine a cluster overview > with cluster aggregate metrics above and a list of regionservers sorted by > utilization below; and a regionserver view with process metrics above and a > list of metrics by operation type below, or a list of client connections, or > a list of threads, sorted by utilization, throughput, or latency. > Generically 'htop' is taken but would be distinctive in the HBase context, a > utility org.apache.hadoop.hbase.HTop > No need necessarily for a curses interface. Could be an external monitor with > a web front end as has been discussed before. I do like the idea of a process > that runs in a terminal because I interact with dev and test HBase clusters > exclusively by SSH. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (HBASE-11062) htop
[ https://issues.apache.org/jira/browse/HBASE-11062?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16847866#comment-16847866 ] Andrew Purtell commented on HBASE-11062: +1 for hbtop. Fits with other naming > htop > > > Key: HBASE-11062 > URL: https://issues.apache.org/jira/browse/HBASE-11062 > Project: HBase > Issue Type: New Feature > Components: hbase-operator-tools >Reporter: Andrew Purtell >Assignee: Toshihiro Suzuki >Priority: Major > > A top-like monitor could be useful for testing, debugging, operations of > clusters of moderate size, and possibly for diagnosing issues in large > clusters. > Consider a curses interface like the one presented by atop > (http://www.atoptool.nl/images/screenshots/genericw.png) - with aggregate > metrics collected over a monitoring interval in the upper portion of the > pane, and a listing of discrete measurements sorted and filtered by various > criteria in the bottom part of the pane. One might imagine a cluster overview > with cluster aggregate metrics above and a list of regionservers sorted by > utilization below; and a regionserver view with process metrics above and a > list of metrics by operation type below, or a list of client connections, or > a list of threads, sorted by utilization, throughput, or latency. > Generically 'htop' is taken but would be distinctive in the HBase context, a > utility org.apache.hadoop.hbase.HTop > No need necessarily for a curses interface. Could be an external monitor with > a web front end as has been discussed before. I do like the idea of a process > that runs in a terminal because I interact with dev and test HBase clusters > exclusively by SSH. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (HBASE-11062) htop
[ https://issues.apache.org/jira/browse/HBASE-11062?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16847865#comment-16847865 ] Josh Elser commented on HBASE-11062: Coming here after Toshi's demo IRL at NoSQL day – I'd echo JMS' warning about confusion with [https://hisham.hm/htop/]. Author's choice, of course, but I'd throw out "hbtop" as a name instead. You have a patch/PR for what you demo'ed (even with the LGPL dependency)? > htop > > > Key: HBASE-11062 > URL: https://issues.apache.org/jira/browse/HBASE-11062 > Project: HBase > Issue Type: New Feature > Components: hbase-operator-tools >Reporter: Andrew Purtell >Assignee: Toshihiro Suzuki >Priority: Major > > A top-like monitor could be useful for testing, debugging, operations of > clusters of moderate size, and possibly for diagnosing issues in large > clusters. > Consider a curses interface like the one presented by atop > (http://www.atoptool.nl/images/screenshots/genericw.png) - with aggregate > metrics collected over a monitoring interval in the upper portion of the > pane, and a listing of discrete measurements sorted and filtered by various > criteria in the bottom part of the pane. One might imagine a cluster overview > with cluster aggregate metrics above and a list of regionservers sorted by > utilization below; and a regionserver view with process metrics above and a > list of metrics by operation type below, or a list of client connections, or > a list of threads, sorted by utilization, throughput, or latency. > Generically 'htop' is taken but would be distinctive in the HBase context, a > utility org.apache.hadoop.hbase.HTop > No need necessarily for a curses interface. Could be an external monitor with > a web front end as has been discussed before. I do like the idea of a process > that runs in a terminal because I interact with dev and test HBase clusters > exclusively by SSH. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (HBASE-11062) htop
[ https://issues.apache.org/jira/browse/HBASE-11062?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16797724#comment-16797724 ] Toshihiro Suzuki commented on HBASE-11062: -- I will work on this tool on hbase-operator-tools repository. https://github.com/apache/hbase-operator-tools > htop > > > Key: HBASE-11062 > URL: https://issues.apache.org/jira/browse/HBASE-11062 > Project: HBase > Issue Type: New Feature >Reporter: Andrew Purtell >Assignee: Toshihiro Suzuki >Priority: Major > Labels: beginner > > A top-like monitor could be useful for testing, debugging, operations of > clusters of moderate size, and possibly for diagnosing issues in large > clusters. > Consider a curses interface like the one presented by atop > (http://www.atoptool.nl/images/screenshots/genericw.png) - with aggregate > metrics collected over a monitoring interval in the upper portion of the > pane, and a listing of discrete measurements sorted and filtered by various > criteria in the bottom part of the pane. One might imagine a cluster overview > with cluster aggregate metrics above and a list of regionservers sorted by > utilization below; and a regionserver view with process metrics above and a > list of metrics by operation type below, or a list of client connections, or > a list of threads, sorted by utilization, throughput, or latency. > Generically 'htop' is taken but would be distinctive in the HBase context, a > utility org.apache.hadoop.hbase.HTop > No need necessarily for a curses interface. Could be an external monitor with > a web front end as has been discussed before. I do like the idea of a process > that runs in a terminal because I interact with dev and test HBase clusters > exclusively by SSH. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (HBASE-11062) htop
[ https://issues.apache.org/jira/browse/HBASE-11062?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16772485#comment-16772485 ] Toshihiro Suzuki commented on HBASE-11062: -- Thank you for the information [~apurtell]. Will work on this. > htop > > > Key: HBASE-11062 > URL: https://issues.apache.org/jira/browse/HBASE-11062 > Project: HBase > Issue Type: New Feature >Reporter: Andrew Purtell >Priority: Major > Labels: beginner > > A top-like monitor could be useful for testing, debugging, operations of > clusters of moderate size, and possibly for diagnosing issues in large > clusters. > Consider a curses interface like the one presented by atop > (http://www.atoptool.nl/images/screenshots/genericw.png) - with aggregate > metrics collected over a monitoring interval in the upper portion of the > pane, and a listing of discrete measurements sorted and filtered by various > criteria in the bottom part of the pane. One might imagine a cluster overview > with cluster aggregate metrics above and a list of regionservers sorted by > utilization below; and a regionserver view with process metrics above and a > list of metrics by operation type below, or a list of client connections, or > a list of threads, sorted by utilization, throughput, or latency. > Generically 'htop' is taken but would be distinctive in the HBase context, a > utility org.apache.hadoop.hbase.HTop > No need necessarily for a curses interface. Could be an external monitor with > a web front end as has been discussed before. I do like the idea of a process > that runs in a terminal because I interact with dev and test HBase clusters > exclusively by SSH. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (HBASE-11062) htop
[ https://issues.apache.org/jira/browse/HBASE-11062?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16770622#comment-16770622 ] Andrew Purtell commented on HBASE-11062: Sure [~brfrn169], please go ahead. We were going to work on this, but then there were shifting priorities. Some fully featured external metrics database and query interface is more than sufficient for the task. If you are large scale, you have something like this, and then there is no pressing need to build something into HBase; and yet, there should probably be something intrinsic available for smaller scale or testers. An earlier interface specification can be found here: https://issues.apache.org/jira/browse/HBASE-11062?focusedCommentId=14707090=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-14707090 > htop > > > Key: HBASE-11062 > URL: https://issues.apache.org/jira/browse/HBASE-11062 > Project: HBase > Issue Type: New Feature >Reporter: Andrew Purtell >Priority: Major > Labels: beginner > > A top-like monitor could be useful for testing, debugging, operations of > clusters of moderate size, and possibly for diagnosing issues in large > clusters. > Consider a curses interface like the one presented by atop > (http://www.atoptool.nl/images/screenshots/genericw.png) - with aggregate > metrics collected over a monitoring interval in the upper portion of the > pane, and a listing of discrete measurements sorted and filtered by various > criteria in the bottom part of the pane. One might imagine a cluster overview > with cluster aggregate metrics above and a list of regionservers sorted by > utilization below; and a regionserver view with process metrics above and a > list of metrics by operation type below, or a list of client connections, or > a list of threads, sorted by utilization, throughput, or latency. > Generically 'htop' is taken but would be distinctive in the HBase context, a > utility org.apache.hadoop.hbase.HTop > No need necessarily for a curses interface. Could be an external monitor with > a web front end as has been discussed before. I do like the idea of a process > that runs in a terminal because I interact with dev and test HBase clusters > exclusively by SSH. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (HBASE-11062) htop
[ https://issues.apache.org/jira/browse/HBASE-11062?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16770403#comment-16770403 ] Toshihiro Suzuki commented on HBASE-11062: -- Can I take this? This is a very good idea. > htop > > > Key: HBASE-11062 > URL: https://issues.apache.org/jira/browse/HBASE-11062 > Project: HBase > Issue Type: New Feature >Reporter: Andrew Purtell >Assignee: Andrew Purtell >Priority: Major > Labels: beginner > > A top-like monitor could be useful for testing, debugging, operations of > clusters of moderate size, and possibly for diagnosing issues in large > clusters. > Consider a curses interface like the one presented by atop > (http://www.atoptool.nl/images/screenshots/genericw.png) - with aggregate > metrics collected over a monitoring interval in the upper portion of the > pane, and a listing of discrete measurements sorted and filtered by various > criteria in the bottom part of the pane. One might imagine a cluster overview > with cluster aggregate metrics above and a list of regionservers sorted by > utilization below; and a regionserver view with process metrics above and a > list of metrics by operation type below, or a list of client connections, or > a list of threads, sorted by utilization, throughput, or latency. > Generically 'htop' is taken but would be distinctive in the HBase context, a > utility org.apache.hadoop.hbase.HTop > No need necessarily for a curses interface. Could be an external monitor with > a web front end as has been discussed before. I do like the idea of a process > that runs in a terminal because I interact with dev and test HBase clusters > exclusively by SSH. -- This message was sent by Atlassian JIRA (v7.6.3#76005)
[jira] [Commented] (HBASE-11062) htop
[ https://issues.apache.org/jira/browse/HBASE-11062?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16314856#comment-16314856 ] stack commented on HBASE-11062: --- Need this. Currently scraping and dumping to tsdb. Next was how to integrate. You are ahead. Was trying to figure list of questions report would answer. Histograms. What you seeing for load? First questions we had were whether scan or random get (if latter then the indexed block fix would help) and then source of loading. Good stuff andy > htop > > > Key: HBASE-11062 > URL: https://issues.apache.org/jira/browse/HBASE-11062 > Project: HBase > Issue Type: New Feature >Reporter: Andrew Purtell >Assignee: Andrew Purtell > Labels: beginner > > A top-like monitor could be useful for testing, debugging, operations of > clusters of moderate size, and possibly for diagnosing issues in large > clusters. > Consider a curses interface like the one presented by atop > (http://www.atoptool.nl/images/screenshots/genericw.png) - with aggregate > metrics collected over a monitoring interval in the upper portion of the > pane, and a listing of discrete measurements sorted and filtered by various > criteria in the bottom part of the pane. One might imagine a cluster overview > with cluster aggregate metrics above and a list of regionservers sorted by > utilization below; and a regionserver view with process metrics above and a > list of metrics by operation type below, or a list of client connections, or > a list of threads, sorted by utilization, throughput, or latency. > Generically 'htop' is taken but would be distinctive in the HBase context, a > utility org.apache.hadoop.hbase.HTop > No need necessarily for a curses interface. Could be an external monitor with > a web front end as has been discussed before. I do like the idea of a process > that runs in a terminal because I interact with dev and test HBase clusters > exclusively by SSH. -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Commented] (HBASE-11062) htop
[ https://issues.apache.org/jira/browse/HBASE-11062?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16314835#comment-16314835 ] Andrew Purtell commented on HBASE-11062: While debugging elevated meta query rates we've noticed how clunky it is to do that right now. In order to find top-N accesses by client identity or by table name you have to toggle RpcServer to TRACE (only for a short time!), scrape the resulting log output, extract via regex, rinse and repeat. It becomes apparent a special set of metrics for the meta region, counting queries for particular tables and by client identity, would be useful. Although this JIRA is scoped to be a lot more than this, meta query statistics would be a good start. This doesn't require any cross cluster (partial) aggregation and communication. Whatever server hosting meta can create a special mbean for tracking meta request stats. This information would be picked up by whatever monitoring solution is already in place. I'll file a subtask for this. > htop > > > Key: HBASE-11062 > URL: https://issues.apache.org/jira/browse/HBASE-11062 > Project: HBase > Issue Type: New Feature >Reporter: Andrew Purtell >Assignee: Andrew Purtell > Labels: beginner > > A top-like monitor could be useful for testing, debugging, operations of > clusters of moderate size, and possibly for diagnosing issues in large > clusters. > Consider a curses interface like the one presented by atop > (http://www.atoptool.nl/images/screenshots/genericw.png) - with aggregate > metrics collected over a monitoring interval in the upper portion of the > pane, and a listing of discrete measurements sorted and filtered by various > criteria in the bottom part of the pane. One might imagine a cluster overview > with cluster aggregate metrics above and a list of regionservers sorted by > utilization below; and a regionserver view with process metrics above and a > list of metrics by operation type below, or a list of client connections, or > a list of threads, sorted by utilization, throughput, or latency. > Generically 'htop' is taken but would be distinctive in the HBase context, a > utility org.apache.hadoop.hbase.HTop > No need necessarily for a curses interface. Could be an external monitor with > a web front end as has been discussed before. I do like the idea of a process > that runs in a terminal because I interact with dev and test HBase clusters > exclusively by SSH. -- This message was sent by Atlassian JIRA (v6.4.14#64029)
[jira] [Commented] (HBASE-11062) htop
[ https://issues.apache.org/jira/browse/HBASE-11062?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14361991#comment-14361991 ] Andrew Purtell commented on HBASE-11062: We should probably pick this back up and ship it with releases now that Cassandra has https://github.com/hailocab/ctop/blob/master/README.md htop Key: HBASE-11062 URL: https://issues.apache.org/jira/browse/HBASE-11062 Project: HBase Issue Type: New Feature Reporter: Andrew Purtell A top-like monitor could be useful for testing, debugging, operations of clusters of moderate size, and possibly for diagnosing issues in large clusters. Consider a curses interface like the one presented by atop (http://www.atoptool.nl/images/screenshots/genericw.png) - with aggregate metrics collected over a monitoring interval in the upper portion of the pane, and a listing of discrete measurements sorted and filtered by various criteria in the bottom part of the pane. One might imagine a cluster overview with cluster aggregate metrics above and a list of regionservers sorted by utilization below; and a regionserver view with process metrics above and a list of metrics by operation type below, or a list of client connections, or a list of threads, sorted by utilization, throughput, or latency. Generically 'htop' is taken but would be distinctive in the HBase context, a utility org.apache.hadoop.hbase.HTop No need necessarily for a curses interface. Could be an external monitor with a web front end as has been discussed before. I do like the idea of a process that runs in a terminal because I interact with dev and test HBase clusters exclusively by SSH. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (HBASE-11062) htop
[ https://issues.apache.org/jira/browse/HBASE-11062?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13979134#comment-13979134 ] Jean-Marc Spaggiari commented on HBASE-11062: - HTop make me think about htop ;) http://hisham.hm/htop/htop-1.0.2-io.png https://packages.debian.org/search?keywords=htop But I agree that it might be nice to have! Like a status page with all those metrics... Cassandra has nice metrics reports already. htop Key: HBASE-11062 URL: https://issues.apache.org/jira/browse/HBASE-11062 Project: HBase Issue Type: New Feature Reporter: Andrew Purtell A top-like monitor could be useful for testing, debugging, operations of clusters of moderate size, and possibly for diagnosing issues in large clusters. Consider a curses interface like the one presented by atop (http://www.atoptool.nl/images/screenshots/genericw.png) - with aggregate metrics collected over a monitoring interval in the upper portion of the pane, and a listing of discrete measurements sorted and filtered by various criteria in the bottom part of the pane. One might imagine a cluster overview with cluster aggregate metrics above and a list of regionservers sorted by utilization below; and a regionserver view with process metrics above and a list of metrics by operation type below, or a list of client connections, or a list of threads, sorted by utilization, throughput, or latency. Generically 'htop' is taken but would be distinctive in the HBase context, a utility org.apache.hadoop.hbase.HTop No need necessarily for a curses interface. Could be an external monitor with a web front end as has been discussed before. I do like the idea of a process that runs in a terminal because I interact with dev and test HBase clusters exclusively by SSH. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (HBASE-11062) htop
[ https://issues.apache.org/jira/browse/HBASE-11062?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13979138#comment-13979138 ] Esteban Gutierrez commented on HBASE-11062: --- I'd been thinking about this for a while and what I'd like to see is something between {{jvmtop}} https://code.google.com/p/jvmtop/, {{/jmx}} metrics and HTrace. There is a lot of useful stuff from the DNs and hosts health that could be used too. But initially {{jvmtop}} functionality to detect offending threads would be really helpful. htop Key: HBASE-11062 URL: https://issues.apache.org/jira/browse/HBASE-11062 Project: HBase Issue Type: New Feature Reporter: Andrew Purtell A top-like monitor could be useful for testing, debugging, operations of clusters of moderate size, and possibly for diagnosing issues in large clusters. Consider a curses interface like the one presented by atop (http://www.atoptool.nl/images/screenshots/genericw.png) - with aggregate metrics collected over a monitoring interval in the upper portion of the pane, and a listing of discrete measurements sorted and filtered by various criteria in the bottom part of the pane. One might imagine a cluster overview with cluster aggregate metrics above and a list of regionservers sorted by utilization below; and a regionserver view with process metrics above and a list of metrics by operation type below, or a list of client connections, or a list of threads, sorted by utilization, throughput, or latency. Generically 'htop' is taken but would be distinctive in the HBase context, a utility org.apache.hadoop.hbase.HTop No need necessarily for a curses interface. Could be an external monitor with a web front end as has been discussed before. I do like the idea of a process that runs in a terminal because I interact with dev and test HBase clusters exclusively by SSH. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (HBASE-11062) htop
[ https://issues.apache.org/jira/browse/HBASE-11062?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13979140#comment-13979140 ] Andrew Purtell commented on HBASE-11062: Too bad jvmtop is GPL, but that seems like a good place to start, combined with jmx metrics. htop Key: HBASE-11062 URL: https://issues.apache.org/jira/browse/HBASE-11062 Project: HBase Issue Type: New Feature Reporter: Andrew Purtell A top-like monitor could be useful for testing, debugging, operations of clusters of moderate size, and possibly for diagnosing issues in large clusters. Consider a curses interface like the one presented by atop (http://www.atoptool.nl/images/screenshots/genericw.png) - with aggregate metrics collected over a monitoring interval in the upper portion of the pane, and a listing of discrete measurements sorted and filtered by various criteria in the bottom part of the pane. One might imagine a cluster overview with cluster aggregate metrics above and a list of regionservers sorted by utilization below; and a regionserver view with process metrics above and a list of metrics by operation type below, or a list of client connections, or a list of threads, sorted by utilization, throughput, or latency. Generically 'htop' is taken but would be distinctive in the HBase context, a utility org.apache.hadoop.hbase.HTop No need necessarily for a curses interface. Could be an external monitor with a web front end as has been discussed before. I do like the idea of a process that runs in a terminal because I interact with dev and test HBase clusters exclusively by SSH. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (HBASE-11062) htop
[ https://issues.apache.org/jira/browse/HBASE-11062?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13979234#comment-13979234 ] Lars Hofhansl commented on HBASE-11062: --- +1000 :) htop Key: HBASE-11062 URL: https://issues.apache.org/jira/browse/HBASE-11062 Project: HBase Issue Type: New Feature Reporter: Andrew Purtell A top-like monitor could be useful for testing, debugging, operations of clusters of moderate size, and possibly for diagnosing issues in large clusters. Consider a curses interface like the one presented by atop (http://www.atoptool.nl/images/screenshots/genericw.png) - with aggregate metrics collected over a monitoring interval in the upper portion of the pane, and a listing of discrete measurements sorted and filtered by various criteria in the bottom part of the pane. One might imagine a cluster overview with cluster aggregate metrics above and a list of regionservers sorted by utilization below; and a regionserver view with process metrics above and a list of metrics by operation type below, or a list of client connections, or a list of threads, sorted by utilization, throughput, or latency. Generically 'htop' is taken but would be distinctive in the HBase context, a utility org.apache.hadoop.hbase.HTop No need necessarily for a curses interface. Could be an external monitor with a web front end as has been discussed before. I do like the idea of a process that runs in a terminal because I interact with dev and test HBase clusters exclusively by SSH. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (HBASE-11062) htop
[ https://issues.apache.org/jira/browse/HBASE-11062?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13979270#comment-13979270 ] stack commented on HBASE-11062: --- Suggest we do Andy's later suggestion, do this as a screen in external app or in master first and then figure how to get it working in a terminal console? (links/lynx?) A basis could be the federation of jmx metrics, or better, a rolling window of jmx readings (1m, 5m, 15m). Or we'd keep metrics in a table for tools like this to read; this would probably be the better way to go. JMX was supposed to have a federated bean view which would have been nice here but seems stalled, jmx 2.0 is stalled... so maybe trying to get a view on cluster via jmx is non-starter and we should go another route.. jmx emissions are fine for a tcollector and its ilk but for an aggregated view in master or in an htop console, it seems like a bunch of custom work -- and jmx w/ its reliance on rmi is probably just a no-go. htop Key: HBASE-11062 URL: https://issues.apache.org/jira/browse/HBASE-11062 Project: HBase Issue Type: New Feature Reporter: Andrew Purtell A top-like monitor could be useful for testing, debugging, operations of clusters of moderate size, and possibly for diagnosing issues in large clusters. Consider a curses interface like the one presented by atop (http://www.atoptool.nl/images/screenshots/genericw.png) - with aggregate metrics collected over a monitoring interval in the upper portion of the pane, and a listing of discrete measurements sorted and filtered by various criteria in the bottom part of the pane. One might imagine a cluster overview with cluster aggregate metrics above and a list of regionservers sorted by utilization below; and a regionserver view with process metrics above and a list of metrics by operation type below, or a list of client connections, or a list of threads, sorted by utilization, throughput, or latency. Generically 'htop' is taken but would be distinctive in the HBase context, a utility org.apache.hadoop.hbase.HTop No need necessarily for a curses interface. Could be an external monitor with a web front end as has been discussed before. I do like the idea of a process that runs in a terminal because I interact with dev and test HBase clusters exclusively by SSH. -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (HBASE-11062) htop
[ https://issues.apache.org/jira/browse/HBASE-11062?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13979279#comment-13979279 ] Andrew Purtell commented on HBASE-11062: As long as we can get a cluster top view that lets you drill down into per server top views. I might tunnel a web interface for that. If the interface can render JSON instead of HTML then additional things like a terminal app can still happen later. htop Key: HBASE-11062 URL: https://issues.apache.org/jira/browse/HBASE-11062 Project: HBase Issue Type: New Feature Reporter: Andrew Purtell A top-like monitor could be useful for testing, debugging, operations of clusters of moderate size, and possibly for diagnosing issues in large clusters. Consider a curses interface like the one presented by atop (http://www.atoptool.nl/images/screenshots/genericw.png) - with aggregate metrics collected over a monitoring interval in the upper portion of the pane, and a listing of discrete measurements sorted and filtered by various criteria in the bottom part of the pane. One might imagine a cluster overview with cluster aggregate metrics above and a list of regionservers sorted by utilization below; and a regionserver view with process metrics above and a list of metrics by operation type below, or a list of client connections, or a list of threads, sorted by utilization, throughput, or latency. Generically 'htop' is taken but would be distinctive in the HBase context, a utility org.apache.hadoop.hbase.HTop No need necessarily for a curses interface. Could be an external monitor with a web front end as has been discussed before. I do like the idea of a process that runs in a terminal because I interact with dev and test HBase clusters exclusively by SSH. -- This message was sent by Atlassian JIRA (v6.2#6252)