<div dir="ltr">Hey All, <div><br></div><div><a href="https://github.com/erlang/otp/pull/161">https://github.com/erlang/otp/pull/161</a><br><div><br></div><div>This patch adds a new option to the erlang:system_monitor BIF, allowing to subscribe monitor process to the long_message_queue events. It should be really helpful when investigating architectural bottlenecks, as well as a part of regular logging procedures. </div>
<div><br></div><div>Few comments about the implementation:</div><div>- Monitor messages are generated once the message gets appended to the receiver's message queue. </div><div>- For a certain process - monitor message can be generated only once after exceeding a pre-configured threshold. Erlang VM will resend the monitor message only once the process mailbox size gets below the limit and exceeds them again.</div>
<div>- Info list of the monitor message currently contains only receiver's message queue length. If you have any ideas on what might be useful to put there in addition to the mailbox size - please let me know.</div><div>
<div><font color="#1155cc" face="arial, sans-serif"><u><br></u></font></div><div><font face="arial, sans-serif">Thanks, </font></div></div><div><font face="arial, sans-serif">Michal </font></div></div></div>