Hi Folks,
I’ve got an odd issue, I’m not sure if this is something I’ve done wrong or an issue with Zabbix. I’m using version 4.4.0.
I’ve built a set of web scenarios in a template which just looks at homepages of some of our company applications. I’ve applied this template to a set of ‘probe’ servers I’ve set up at different sites across the globe. The probe servers have Zabbix agents running.
All seemed to be working fine until I noticed that the response times for these probes were the same, when there should be significant variation between them. An HTTP call from India to UK should obviously be much slower than one from a machine in the same UK data centre the target application server is in.
After double checking the config and building a web scenario directly on a probe rather than using a template (just in case) I still found the response times to be unrealistically fast. I started testing curls from the probe and found them to be much slower to the same target server than the web scenarios.
I then put tshark on a couple of the servers and put a filter in place to detect the raw traffic coming from the probe server to the target server. No traffic seen over the course of several web scenario runs. After further thought, I put tshark on the Zabbix server and found the missing traffic, all the web scenarios were running on the Zabbix server. Every time I bring up a new host it adds another set of web calls being made from the Zabbix server.
My question is, is this something which I’ve set up wrong, or is this by design? Does Zabbix only allow you to run web scenarios from the Zabbix server?
Thanks in advance for any forthcoming help!
I’ve got an odd issue, I’m not sure if this is something I’ve done wrong or an issue with Zabbix. I’m using version 4.4.0.
I’ve built a set of web scenarios in a template which just looks at homepages of some of our company applications. I’ve applied this template to a set of ‘probe’ servers I’ve set up at different sites across the globe. The probe servers have Zabbix agents running.
All seemed to be working fine until I noticed that the response times for these probes were the same, when there should be significant variation between them. An HTTP call from India to UK should obviously be much slower than one from a machine in the same UK data centre the target application server is in.
After double checking the config and building a web scenario directly on a probe rather than using a template (just in case) I still found the response times to be unrealistically fast. I started testing curls from the probe and found them to be much slower to the same target server than the web scenarios.
I then put tshark on a couple of the servers and put a filter in place to detect the raw traffic coming from the probe server to the target server. No traffic seen over the course of several web scenario runs. After further thought, I put tshark on the Zabbix server and found the missing traffic, all the web scenarios were running on the Zabbix server. Every time I bring up a new host it adds another set of web calls being made from the Zabbix server.
My question is, is this something which I’ve set up wrong, or is this by design? Does Zabbix only allow you to run web scenarios from the Zabbix server?
Thanks in advance for any forthcoming help!
Comment