The ewgapagent process is a component of the SeisNetWatch that constantly polls an Earthworm Wave Ring for gaps and overlaps of data. SeisNetWatch is a parametric rule-based network monitoring system comprised of agents (which feed data), an Network Station Information (NSI) server (which collects and interprets data), and clients (which obtain information from the server). The ewgapagent is a parameteric data feeding agent that observes trace_buf packets flowing through an Earthworm Ring Buffer and reports what it sees to the NSI server. The primary information that the ewgapagent reports is that trace data for any channel, for stations at this installation, that are showing gaps or overlaps.
The ewgapagent is a product of ISTI. Contact us for more information (and licensing if you are interested in using this product).
After reading in all of its parameters from the configuration file, the ewgapagent attaches to the Event Channels running on the same server where the NSI is located. The ewgapagent will wait for the Event Channels to appear before doing any polling. Should any of these Event Channels disappear (because the NSI's host dies), the ewgapagent will exit.
Once the agent attaches to the Event Channels successfully, it attaches to the WaveRing specified and begins a polling cycle. The polling cycle for the agent will run until the program is terminated.
The ewgapagent polls the wave ring constantly and reports when it finds gaps or overlaps between packets for the same station and channel.
The user has the ability to ignore certain channels of data by specifying the first 1 or 2 chars of the channel name in a configuration parameter. This is useful for ignoring triggered channels.
It dynamically stores the SCN for each packet it observes and measures the following values for later reporting:
In addition, the ewgapagent reports which channel reported a gap, when the gap analysis was started, and the time of the last gap/overlap report.
ISTI recommends monitoring the 24 hour cumulative gap for various channels and using these as voting parameters within SeisNetWatch to detect stations that have excessive gaps and timing problems.