[ZBX-13616] As Trapper processes adjust monitoring time, trigger evaluations use incorrect data Created: 2018 Mar 15  Updated: 2018 Mar 15  Resolved: 2018 Mar 15

Status: Closed
Project: ZABBIX BUGS AND ISSUES
Component/s: Proxy (P), Server (S)
Affects Version/s: 2.2.22rc1, 3.0.15
Fix Version/s: None

Type: Incident report Priority: Trivial
Reporter: MATSUDA Daiki Assignee: Unassigned
Resolution: Duplicate Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Issue Links:
Duplicate
duplicates ZBXNEXT-3298 Keep track of per-host time differences Open
duplicates ZBX-12390 Illegal trigger detection. Closed
duplicates ZBX-12957 zabbix agentd use active mode traffic... Closed

 Description   

Zabbix Server and Proxy Trapper processes receive data from Agents and Proxies. Then they adjust monitoring time with included timestamp.

The data from Agent, especially log or eventlog, is sent by big block. But from Proxy, is sliced in spite of reading log almost same time. By this, in a
spcific situation, data timestamp is reverted.
e.g.

Agent reading
data 1 : time a
data 2 : time b := a + 0.001
send data 1 and 2 together

Proxy receiving
data 1 : time c := a + 0.01
data 2 : time d := b + 0.01 := a + 0.011

Send data 1 and 2 separately

Server receving
data 1 : time e := c + 0.02 := a + 0.03
data 2 : time f := d + 0.01 := a + 0.021

The revert occures.

In addition, the trigger with regexp() selects data with the accuracy second not nano second. So, when trigger evaluates against data 2, select data 1 because it selects the latest data.



 Comments   
Comment by Glebs Ivanovskis (Inactive) [ 2018 Mar 15 ]

Good news: timestamp corrections will be removed in ZBX-12957, value cache precision will be improved in ZBX-12390, work in progress.

I'm going to close this issue as Duplicate, but anyway, thank you for reporting it!

Generated at Sat Apr 20 13:52:46 EEST 2024 using Jira 9.12.4#9120004-sha1:625303b708afdb767e17cb2838290c41888e9ff0.