This may indicate an SSL negotiation failure or timeout. Running tcpdump
may confirm a timeout.
Packets from an SSL session can be inspected in Wireshark and ssldump, if
you can get access to the private key.
If this happens only a few times a week, you may want to just use
"DELAYRED" (see hosts.cfg) to allow xymonnet to try again a minute later.
J
On 2 July 2018 at 20:53, Jonas Weber <user-aaaab6ddd076@xymon.invalid> wrote:
Hello,
in our Xymon installation we are experiencing red HTTP status 0 alerts
repeatedly, about a few times per week. This phenomenon started appearing
when we switched the backend from Apache to nginx. The monitored service is
checked over an SSL connection, making tcpdump et al. difficult to use.
Additionally, this only appears on our production monitoring; a secondary
installation with only this check does not show this issue. So, debug
logging isn't really feasible, either (although I've tried).
At this point, we are quite certain that it is a monitoring artifact and
not a problem of the service, as there are no errors on the nginx side of
things, and no customers have complained, either.
I've looked in the code of xymonnet. I think status code 0 might mean that
no HTTP/1.x response has been received, or at least that Xymon was unable
to parse it. Is that correct? Does/did anybody else face this problem?
Best regards,
Jonas