On Tue, February 24, 2015 1:47 pm, Troy Adams wrote:
Is there anyway to immediately do a TCP trace when an HTTP service goes down and include that data in the status page?
cheers,
Troy
Troy,
At the moment, no. trace's of any type can only be configured to be automatically performed on a ping/conn test failure. (Using the --trace option to xymonnet and/or the 'trace' option in hosts.cfg.)
Unfortunately, adding it in at the time of the initial status result doesn't scale very well (for any trace at all), since xymonnet can't send its reports in until the trace completes.
It would be nice to be able to complete this afterwards (either as a built-in feature, or via a custom script) and be able to add that into an existing status message in a way that causes another snapshot to be taken, but the current model doesn't really support that.
One possible integration point would be a modify message, however they can only easily be on a single line, and one re-iterating the 'red' color status, while it'll be displayed, won't cause a new snapshot to be taken recording the trace itself for posterity.
If the record of the trace after the situation is resolved isn't important to keep, that might be an option.
HTH,
-jc