Please clarify: http yellow when delayred is set? Or, what makes http yellow?
list Betsy Schwartz
I have this default set: 0.0.0.0 .default. # DOWNTIME=0:0000:0300 delayred=http:10 and servers that are monitoring three URL's like so: 10.1.2.3 web1.example.com # CLASS:webzone ssh apache http://web1.example.com/url1.htm http://web1.example.com/url2.cfm http://web1.example.com/url3.cfm What I'm seeing is, with one URL going red with a timeout after 22 seconds, the http test is going YELLOW and then going RED ten minutes later. Does delayred turn the test yellow? If not, why is this test turning yellow? We want it to be RED. What exactly determines whether an http test is red or yellow? thanks Betsy
list Betsy Schwartz
Any thought on this? We just had another missed outage. I am going to have to start paging on http yellow until this is resolved - UGH Thanks Betsy
▸
On Jun 7, 2012, at 8:48 AM, Betsy Schwartz <user-c61747246f66@xymon.invalid> wrote:
I have this default set: 0.0.0.0 .default. # DOWNTIME=0:0000:0300 delayred=http:10 and servers that are monitoring three URL's like so: 10.1.2.3 web1.example.com # CLASS:webzone ssh apache http://web1.example.com/url1.htm http://web1.example.com/url2.cfm http://web1.example.com/url3.cfm What I'm seeing is, with one URL going red with a timeout after 22 seconds, the http test is going YELLOW and then going RED ten minutes later. Does delayred turn the test yellow? If not, why is this test turning yellow? We want it to be RED. What exactly determines whether an http test is red or yellow? thanks Betsy
list Ralph Mitchell
I don't know why it's flagging the way it is, but could you break it out
into individual checks?
10.1.2.3 web1.example.com # CLASS:webzone ssh apache
http://web1.example.com/url1.htm
0.0.0.0 web1-url2 # noconn
http://web1.example.com/url2.cfm
0.0.0.0 web1-url3 # noconn
http://web1.example.com/url3.cfm
It's not ideal, but it would give you separate visibility and reporting
until the yellow/red problem is worked out.
Ralph Mitchell
▸
On Wed, Jun 13, 2012 at 6:16 PM, Betsy Schwartz <user-c61747246f66@xymon.invalid>wrote:
Any thought on this? We just had another missed outage. I am going to have to start paging on http yellow until this is resolved - UGH Thanks Betsy On Jun 7, 2012, at 8:48 AM, Betsy Schwartz <user-c61747246f66@xymon.invalid> wrote:I have this default set: 0.0.0.0 .default. # DOWNTIME=0:0000:0300 delayred=http:10 and servers that are monitoring three URL's like so: 10.1.2.3 web1.example.com # CLASS:webzone ssh apache http://web1.example.com/url1.htm http://web1.example.com/url2.cfm http://web1.example.com/url3.cfm What I'm seeing is, with one URL going red with a timeout after 22 seconds, the http test is going YELLOW and then going RED ten minutes later. Does delayred turn the test yellow? If not, why is this test turning yellow? We want it to be RED. What exactly determines whether an http test is red or yellow? thanks Betsy
list Betsy Schwartz
That is a good workaround, thanks. The workaround I chose was to make this particular alert page on yellow. But I really need to clarify my understanding of why an http alert would turn yellow. We have a *lot* of mission-critical http alerts.
▸
On Thu, Jun 14, 2012 at 9:06 PM, Ralph Mitchell <user-00a5e44c48c0@xymon.invalid> wrote:I don't know why it's flagging the way it is, but could you break it out into individual checks? 10.1.2.3 web1.example.com # CLASS:webzone ssh apache http://web1.example.com/url1.htm 0.0.0.0 web1-url2 # noconn http://web1.example.com/url2.cfm 0.0.0.0 web1-url3 # noconn http://web1.example.com/url3.cfm It's not ideal, but it would give you separate visibility and reporting until the yellow/red problem is worked out. Ralph Mitchell
On Wed, Jun 13, 2012 at 6:16 PM, Betsy Schwartz <user-c61747246f66@xymon.invalid>
▸
wrote:Any thought on this? We just had another missed outage. I am going to have to start paging on http yellow until this is resolved - UGH Thanks Betsy On Jun 7, 2012, at 8:48 AM, Betsy Schwartz <user-c61747246f66@xymon.invalid> wrote:I have this default set: 0.0.0.0 .default. # DOWNTIME=0:0000:0300 delayred=http:10 and servers that are monitoring three URL's like so: 10.1.2.3 web1.example.com # CLASS:webzone ssh apache http://web1.example.com/url1.htm http://web1.example.com/url2.cfm http://web1.example.com/url3.cfm What I'm seeing is, with one URL going red with a timeout after 22 seconds, the http test is going YELLOW and then going RED ten minutes later. Does delayred turn the test yellow? If not, why is this test turning yellow? We want it to be RED. What exactly determines whether an http test is red or yellow? thanks Betsy
list Xymon User in Richmond
I'm confused about this, Elizabeth, and must be missing something. Your original setup had the 3 urls on one host line, with delayred=10 set as a default. On a failure, the red was delayed 10 minutes, as configured. It's not clear why the yellow in the interim is a problem, but I'd try using delayyellow=11 if I wanted to leave it green until the red kicked in. Or is it that you want this one not to use the .default delayred and go red immediately? Wouldn't setting delayred=0 on that particular host do that? (The manpage also indicates that delayred should be silently ignored on a .default entry, but what you're describing seems to indicate that it does work) I'm sure that when you tell me where I'm off track, I'll face-palm. regards, j.
▸
On Fri, June 15, 2012 16:33, Betsy Schwartz wrote:That is a good workaround, thanks. The workaround I chose was to make this particular alert page on yellow. But I really need to clarify my understanding of why an http alert would turn yellow. We have a *lot* of mission-critical http alerts. On Thu, Jun 14, 2012 at 9:06 PM, Ralph Mitchell <user-00a5e44c48c0@xymon.invalid> wrote:I don't know why it's flagging the way it is, but could you break it out into individual checks? 10.1.2.3 web1.example.com # CLASS:webzone ssh apache http://web1.example.com/url1.htm 0.0.0.0 web1-url2 # noconn http://web1.example.com/url2.cfm 0.0.0.0 web1-url3 # noconn http://web1.example.com/url3.cfm It's not ideal, but it would give you separate visibility and reporting until the yellow/red problem is worked out. Ralph Mitchell On Wed, Jun 13, 2012 at 6:16 PM, Betsy Schwartz <user-c61747246f66@xymon.invalid> wrote:Any thought on this? We just had another missed outage. I am going to have to start paging on http yellow until this is resolved - UGH Thanks Betsy On Jun 7, 2012, at 8:48 AM, Betsy Schwartz <user-c61747246f66@xymon.invalid> wrote:I have this default set: 0.0.0.0 .default. # DOWNTIME=0:0000:0300 delayred=http:10 and servers that are monitoring three URL's like so: 10.1.2.3 web1.example.com # CLASS:webzone ssh apache http://web1.example.com/url1.htm http://web1.example.com/url2.cfm http://web1.example.com/url3.cfm What I'm seeing is, with one URL going red with a timeout after 22 seconds, the http test is going YELLOW and then going RED ten minutes later. Does delayred turn the test yellow? If not, why is this test turning yellow? We want it to be RED. What exactly determines whether an http test is red or yellow? thanks Betsy
list Betsy Schwartz
I'm sorry, that was my fault. I've got multiple servers that are each
monitoring multiple URL's, and I'm seeing different behavior on
different servers, just to make life complicated.
All of the servers have two or three URL's monitored.
Some of the time, http goes yellow and then goes red after 10 minutes,
when one URL out of three is unreachable
Most of the time, though, http is just going yellow and staying yellow
when one URL out of three is unreachable.
I claim http shouldn't go yellow **at all**. But I would like a clear
definition of how it's supposed to behave.
The URL's that are failing are Cold Fusion URL's and appear in the
history log as , for example:
red http://host.example.com/monitor/monitor.cfm - Server timeout
Seconds: 22.01