Following up on this for closure, I ended up doing something which included dependencies.
1) Created a check for the DNS lookup, with a noalert. It simply succeeds or fails on whether the DNS lookup works. 2) Created the actual checking procedure, which has a "depends" on #1. I had to add a "for 3 cycles" on #2 because it looks like Monit doesn't actually execute the process checks in an order based on depends statements. Also, there looks to be a one check lag between the actual results and what Monit reports. So the number 3 allowed me to effectively "pause" the checks for #2 based on whether the DNS was working. Thanks for all the help! On Thu, Mar 13, 2025 at 3:08 AM Tino Hendricks via This is the general mailing list for monit <[email protected]> wrote: > Hi Steven, > > the only thing coming to my mind if you can’t handle this in the script > you’re using is > > https://mmonit.com/monit/documentation/monit.html#SERVICE-DEPENDENCIES > > Like, depending on a test for working DNS. > > Cheers! > > Tino > > Am 13.03.2025 um 05:23 schrieb Steven Christensen <[email protected]>: > > > Hi All, I have a use case I am not sure Monit can handle. I looked at the > document and tried keywords in this mailing list, but nothing came up. > > I have a script which checks another node for some information, and then > reports success or failure (via exit code).The script works fine in Monit. > > But due to some dodgy DNS service, sometimes when the script tries to get > information from the other node, the DNS lookup fails. So we don't know > whether the information would have resulted in a success or failure, to > report to Monit. > > My question: Is there a way to tell Monit during that script check > interval that the results of the check are indeterminate, that it should > maintain the previous success/failure state until it the next time it is > able to definitively report a true or false status? > > Thanks! > >
