Zabbix with API knowledge needed for basic setup

I need some startup help for configuring Zabbix (by API only!). It is not magic but my progress is too slow so if you know Zabbix and its API this job is for you! I just installed - VM with Zabbix server, agent and web-interface - VM with mariadb-DB for the server - VM with Zabbix proxy (active), mariadb and agent - VM with Zabbix proxy (passive), mariadb and agent - a few VMs with agents, which will be accessed by proxy (passive or active) or by the server directly Everything is CentOS 7.X. Zabbix is 4.4.4 Now I want to use the API to configure all the rest (users, groups, hosts, triggers, etc...) Yes, I want to use only the API (OK, the web-setup oder the zabbix_proxy etc do need basic changes in the conf-files. That cannot be done by API.) and no, you will not get access to my system. Now show me how to use Zabbix! First: How has the [login to view URL] and [login to view URL] to look like and what are usable values? (ConfigFrequency=3600 is obviously too long!) What about heartbeat? Is communication encryption between server and proxy really as slow as [login to view URL] says? Is there a better way to encrypt the communcation? What is best practice for identifying a proxy (Really by 'Hostname')? Then the API: Change 'Admin'-Password Configure Zabbix being able to send e-mail Add Hosts, Proxies, items, create groups, users, give users rights to see host groups, and to receive notifications about problems etc... What kinds of grouping/tagging does Zabbix offer? E.g. some machines are a part of the load-balancer-setup. So I want to mark/tag/group them. Is this possible? We have no network discovery! monitor 'the typical' on a linux machine by using the templates (ping to it, CPU, Swap, Mem, IO, Disk etc...) by proxy (active, passive) and direct from server and create graphs for it and send notifications. Are the typical thresholds OK for notifications or should some of these customized? Do some web-tests: Access URLs (from proxies (active, passive) or server) and check for return-code or special content in the body or the answer-speed and graph this. Is it possible to have something like # curl_format='{ "time_namelookup": %{time_namelookup}, "time_connect": %{time_connect}, "time_appconnect (e.g after SSL handshake is done)": %{time_appconnect}, "time_pretransfer (e.g. after negotiations)": %{time_pretransfer}, "time_starttransfer": %{time_starttransfer}, "time_total": %{time_total} }' # curl -sS -w "$curl_format" -o /dev/null [login to view URL] { "time_namelookup": 0.041636, "time_connect": 0.052084, "time_appconnect (e.g after SSL handshake is done)": 0.085941, "time_pretransfer (e.g. after negotiations)": 0.086054, "time_starttransfer": 0.132235, "time_total": 0.133771 } in one run in Zabbix to see where the URL-check does spend it's time? How sophisticated is the web URL-check? Am I able to send special headers with the request? Am I able to determine the used TLS-Version or use a different IP for a hostname (for HTTPS) etc? Or do I have to use an external program for that? Is the standard dashboard usable? (I do not need that stupid clock!) Do you have a better setup? For a start we want to monitor around 20 VMs. You tell me the above 'basics' and I will try to start to monitor them and ask if I have problems..... For acessing I am using curl and/or pure Python. (I didn't try to use [login to view URL] if you think it's worth I will try.) Yes, this project is only for a quick start and yes, everything should be done without a click. If you do not know the API do not bid! If you don't have a local Zabbix for trying do not bid, because you will not get access to my systems! As you see: I am in hurry. If you know Zabbix and the API it's an easy job.... Maybe later we talk about monitoring Java/JMX and other applications, nicer graphics (e.g. by Grafana), more sophisticated setups. But that will be a different project.