On latency in event processing network


 
Thread Tools Search this Thread
Special Forums News, Links, Events and Announcements Complex Event Processing RSS News On latency in event processing network
# 1  
Old 08-05-2008
On latency in event processing network

2008-08-05T09:13:00.005+03:00
Image
Packing, on my way to a family vacation in Anatalya, Turkey.

Following an interesting discussion yesterday about performance metrics -- it turns out that definition of latency in event processing network is quite tricky. The reason is that an event can move in multiple paths over the network, in some it is filterred out, in some it just getting into internal state of agent since it does not complete a pattern, and in some it may complete a pattern and trigger reaction, thus there are various ways to define the metrics here. This is important since optimization has to take into account the goal function - what is being optimized. More discussion on this area -- after I'll return. I'll be out of touch for a week.

Source...
Login or Register to Ask a Question

Previous Thread | Next Thread
Login or Register to Ask a Question
PMEVENT(1)						      General Commands Manual							PMEVENT(1)

NAME
pmevent - report event record details SYNOPSIS
pmevent [-gLz] [-a archive] [-h host] [-K spec] [-O offset] [-p port] [-S starttime] [-s samples] [-T endtime] [-t interval] [-x pattern] [-Z timezone] metricname ... DESCRIPTION
Performance Co-Pilot (PCP) supports event records within the framework for fetching general performance information. pmevent prints cur- rent or archived values for the nominated event record metrics. The event records of interest are contained in one or more of the metrics identified by the metricname arguments. Unless directed to another host by the -h option, or to an archive by the -a option or to a local context by the -L option, pmevent will contact the Performance Metrics Collector Daemon (PMCD) on the local host to obtain the required information. The -a,-h and -L options are mutually exclusive. The metricname arguments may be given in the metric specification syntax, as described in PCPIntro(1), where the source and metric name may all be included in the metricname, e.g. thathost:someagent.event.records or myarchive/someagent.event.records['foo-instance','bar- instance']. When this format is used, any of the -h or -a or -L options may also be specified, provided the usage is consistent in terms of the source of the metrics identified by the options as compared to any explicit source of the metrics defined in the metricname argu- ments. When using the metric specification syntax, the ``hostname'' @ is treated specially and causes pmevent to use a local context to collect metrics from PMDAs on the local host without PMCD (same as the -L option). Only some metrics are available in this mode. The -S, -T and -O options may be used to define a time window to restrict the samples retrieved, set an initial origin within the time win- dow; refer to PCPIntro(1) for a complete description of these options. When processing an archive, pmevent may relinquish its own timing control, and operate as a ``slave'' of a pmtime(1) process that uses a GUI dialog to provide timing control. In this case, either the -g option should be used to start pmevent as the sole slave of a new pmtime(1) instance, or -p should be used to attach pmevent to an existing pmtime(1) instance via the IPC channel identified by the port argument. The other options that control the information reported by pmevent are as follows: -a Performance metric values are retrieved from the PCP archive log file identified by the base name archive. -g Start pmevent as the slave of a new pmtime(1) process for replay of archived performance data using the pmtime(1) graphical user interface. -h Current performance metric values are retrieved from the nominated host machine. -K When fetching metrics from a local context, the -K option may be used to control the DSO PMDAs that should be made accessible. The spec argument conforms to the syntax described in __pmSpecLocalPMDA(3). More than one -K option may be used. -L Causes pmevent to use a local context to collect metrics from PMDAs on the local host without PMCD. Only some metrics are available in this mode. -p Attach pmevent to an existing pmtime(1) time control process instance via the IPC channel identified by the port argument. This option is normally only used by other tools, e.g. pmchart(1), when they launch pmevent with synchronized time control. -s The argument samples defines the number of samples to be retrieved and reported. If samples is 0 or -s is not specified, pmevent will sample and report continuously (in real time mode) or until the end of the PCP archive (in archive mode). It is not possible to control the number of event records, as each value of a metricname may deliver zero, one or more event records. The -s option determines how many times pmevent will retrieve values for the specified metricname metrics. -t The default sampling interval may be set to something other than the default 1 second. The interval argument follows the syntax described in PCPIntro(1), and in the simplest form may be an unsigned integer (the implied units in this case are seconds). For PCP archives, pmevent will retrieve all of the event records for the metricname metrics within the requested time window, so the value of the sampling interval will have no effect in this case. -x The given filter is sent to the performance metric domain agent for the requested metricname before any values are requested. This serves two purposes. Firstly, it provides a mechanism for server-side event filtering that is customisable for individual event streams. In addition, some performance metrics domain agents also use the PMCD store mechanism to provide a basic security model (e.g. for sensitive log files, only a client host with pmStore(3) access would be able to access the event stream). -Z By default, pmevent reports the time of day according to the local timezone on the system where pmevent is run. The -Z option changes the timezone to timezone in the format of the environment variable TZ as described in environ(5). -z Change the reporting timezone to the local timezone at the host that is the source of the performance metrics, as identified via either the metricname or the -h or -a or -L options. The output from pmevent is directed to standard output. SEE ALSO
PCPIntro(1), pmcd(1), pmchart(1), pmdumplog(1), pminfo(1), pmlogger(1), pmtime(1), pmval(1), PMAPI(3), __pmSpecLocalPMDA(3), pcp.conf(5) and pcp.env(5). DIAGNOSTICS
All are generated on standard error and are intended to be self-explanatory. Performance Co-Pilot PMEVENT(1)