We are using Apache web server and dynatrace agents are instrumented on each web server. We observed web server was down but we didn't receive any alerts and observed agents were up & running. My question "Is web server agents instrumentation will be different from App server agents?"
The Web Server Agent has a master process that runs as a service that the child apache processes attach to (sort of like a "mini-collector"). If the apache child process is shutdown normally, there will be no incident raised, unless you enable the (disabled by default) Application Process Shutdown incident. If you do enable this incident, you will get an alert every time a process is shutdown. However, if you make good use of Downtimes, this may work well for you without too many false alarms.
Did you expand the timeframe to find the the alert?
If there is no alert in the 'Incident Dashboard' then check the agent logs because in any case you will have to have the alert there whether you received an email or not.
Which server downtime you are talking about? Are you talking about 'Incident' downtime?
Did your shutdown of apache take place outside of the downtime you defined?
I suspected that maybe the incident would fire only if the master agent was shutdown, but I confirmed (on v6.5.10 running on windows) that it fires for either the master or the child shutting down.
So, if you did a normal shutdown (not a kill of the process) and had no Downtime in effect, I am not sure why it would not fire for you.