[alpha4] just some small "problems"

[studder]

Nice work on the new layout.

I found some things in the alpha4-version that you might find interesting:

First; when using a fixed y-axis, the graph isn't bound to this limit.



Second; when i'm logged-in in config-mode and the session is timed-out, a php-error is displayed.

Third; How do I temporarely disable a server which is temporarely unavailable for Zabbix when Zabbix already marked the server as "unreachable"

Fourth; when an agent locks, the service (10050) seems available but no responses are given ->the queue filles up -> other items get affected. It looks like there is no timeout on server-/suckerd-requests. (Zabbix 1.0 & 1.1a4)

Fifth; I use SNMP to monitor IIS on some Win2k machines (can't use the Zabbix agent ). During the night I stop the IIS-service on those machines to do some "things", thats when the trouble starts: the firewall-admin calls me the next day to tell me that I'm flooding the firewall with SNMP-requests. (Zabbix 1.0)

[Alexei]

Welcome to the forums!

1. How would like like ZABBIX to handle situations when values are greater than the limit?

2. This will be fixed.

3. No way yet. I'm aware of the issue. Will be addressed in 1.1.

4. I doubt ZABBIX didn't timeout when agent is slow or unresponsive. Check LogFile or Host error in the list of hosts (Host Configuration). Unavailability of an agent (actually any timeout situation) obviously makes one ZABBIX process spend time on waiting for responses. I've already designed algorithm which will be protected from such situation. The algorithm will be implemented along with switch to multi-threaded model, and this will allow to retrieve hundreds of metrics simultaneously.

5. What do you mean saying "flooding". ZABBIX is trying to reconnect once per 60 seconds if a host is unreachable, no flooding.