Update interval in web scenario does not work correctly

[Freezstyler]

Hello,
I sucesfully created a web scenario to monitor our systems. Everything, triggers and sending emails works correctly, but the update interval is wrong. Interval is set to 60s (i tried also 1m), but the last check is always in 2m+ interval. I need it to check it more often (60s). As you can see on time stamp, it does not work. I even tried different types of agent and more attemps, but nothing helped.

In steps are three steps, which everyone has 17s to failure (Log in, Login check, redirect to tested webpage).

Any thoughts why the interval is checking in 2 minutes, but not in 1m?

Thank you.

EDIT: I discovered that even with same intervals some items are collecting data in 2 min interval (even if it's set to 1m) and some does it in 4m interval ... It's only in web scenario, for SNMP host work even 5s interval like a charm.

EDIT2: Could it be the problem mentionted in Release Notes for Zabbix 6.2.3? I'm running 6.2.2.

ZBX-21517

Fixed invalid update interval handling in network discovery and web scenario scheduling

[Freezstyler]

Anyone knows? There is some strange behaviour. For Log in i have 10s interval, but I'm getting this kind of messages:

2023-01-12 10:34:16	Timeout was reached: Operation timed out after 4308 milliseconds with 0 bytes received
2023-01-12 10:23:17	Timeout was reached: Operation timed out after 10001 milliseconds with 0 bytes received
2023-01-12 08:06:27	Timeout was reached: Operation timed out after 2740 milliseconds with 0 bytes received

10001 ms is fine, because it's 10 second, but the values 4308 and 2740 are wrong.

[fvilarnovo]

Hello, Is there a specific way to reproduce this issue? I can set web scenarios at 10s on 6.2.5 without any problem. Could be the web application that you are monitoring not responding fast enough?

Another way to test this could be use curl (from the same node) and check if times are consistant with what zabbix is doing. For example:

curl -w "Connect time: %{time_connect}\nTime to first byte: %{time_starttransfer}\nTotal: %{time_total}\n" -o /dev/null -sL http://www.zabbix.com

Connect time: 0.116519
Time to first byte: 0.731797
Total: 0.784743

Hope it helps!

[Freezstyler]

Hello,
I'm not an server administrator unfortunately, so I don't have access to server where zabbix runs. Now I only work on web, so I can't try curl.

I don't think it's the problem in application, cause these times are consistent (every time is 4 min interval) and the page responds in 2s and timeout is set to 10s so there is plenty of time.

There is no way how to reproduce this. It's for our ISP system and I can't make and access for someone. I will try explain how I have set it.

There are four steps and Update interval for scenario is 2m, attemps are set for 1:

1) Log in - timeout 10s, requrired status code 200
2) Log in check - timeout 10s, required string and status code 200
3) Forward to another page - timeout 30s, required string and status code 200
4) Log out - timeout 15s, required string and status code 200

I tried many settings, but could't achieve to check it even in 2 min interval instead of four.

I also tried simple check of google page (https://google.com), interval is set to 1m and it checks in 2-3m interval. So I don't think there will be problem with our ISP system, but in some zabbix setting?

[Freezstyler]

And now it's ever more strange. The Log in step will fail, log says this:

555623:20230209:160256.576 cannot process step "Log in" of web scenario "inquiry-read" on host "STIS web control - separated": Timeout was reached: Operation timed out after 4600 milliseconds with 0 bytes received

There's no reason to fail after 4600ms, timeout is set to 10000ms.

Another strange thing is if it fails after 10000ms with this message: Timeout was reached: Resolving timed out after 10000 milliseconds, then the graph for Log in is fine and it's nowhere near to 10s. The problem was generated at 17:48:02