Timstamp is not matched for 1 interface port in 2 item that related with this port

[sasanaalem]

I have these valuese for 1 port that is incoming and outgoing item
here is incoming

IXP: Incoming traffic on interface port-channel2.1968 (IXP) [5m]
2024-05-31 15:32:15

1401318980

2024-05-31 15:27:15

1406430336

2024-05-31 15:22:15

1440054000

2024-05-31 15:17:15

1405963732

and outgoing

IXP: Outgoing traffic on interface port-channel2.1968 (IXP) [5m]
2024-05-31 15:28:13

6194868488

2024-05-31 15:23:13

6072900456

2024-05-31 15:18:13

5979428512

2024-05-31 15:13:13

6120844048

but as you can see the time stamp is not matched!why and how can i fix it in zabbix!?
my network device is same and its the incoming and outgoing for 1 same port ,the interval is 300s and not problem!

[sasanaalem]

Is there anyone to help?!

[Markku]

When Zabbix server has lots of passive checks (like SNMP queries) to schedule, it should distribute the checks somehow more or less evenly so that there is always something to do, instead of having lots of checks executed at the same time and then having long idle times. But, if bulk SNMP requests are used (that's the normal case), naturally the bulk data would arrive at the same time.

Is your switch host configured to use SNMP bulk requests?

Markku

[sasanaalem]

my SNMP is ok!and I dint lost my connection!I have checked it!
also I just want to sync the time!I have a total graph with calculated type but when I want to get a report the itemids that make this total graph are not synched!and it make me a lot of problems!

[cyber]

Then schedule your checks, if they have to be at exact moments...

[sasanaalem]

Here is my itemids,and also i write

Custom intervals>Scheduling>m0-59/5

https://www.zabbix.com/documentation/3.2/en/manual/config/items/item/custom_intervals

but still they are not sync!
also I use snmp!

SELECT itemid, clock, value FROM history_uint WHERE itemid IN (729401, 729604) ORDER BY clock DESC LIMIT 10;
+--------+------------+-----------+
| itemid | clock | value |
+--------+------------+-----------+
| 729604 | 1718192404 | 208556676 |
| 729401 | 1718192401 | 196170892 |
| 729604 | 1718192400 | 208556676 |
| 729401 | 1718192201 | 207474220 |
| 729604 | 1718192104 | 266904108 |
| 729401 | 1718192101 | 228779812 |
| 729604 | 1718192100 | 266904108 |
| 729401 | 1718191901 | 230899252 |
| 729604 | 1718191804 | 238016928 |
| 729604 | 1718191802 | 238016928 |
+--------+------------+-----------+

[sasanaalem]

Then schedule your checks, if they have to be at exact moments...

Could you please point out any mistakes I made?

[cyber]

You are scheduling your calculated item, but you are not scheduling items which collect data... thus, (I think), you are just chasing the wind on a field here... trying to calculate on a precise moments, when it really does not matter... couple of seconds here and there, when data can come in on a different schedule...

[Markku]

So, how is it, you didn't say: Is your switch host (in Zabbix configuration) configured to use SNMP bulk requests?

Markku

[sasanaalem]

So, how is it, you didn't say: Is your switch host (in Zabbix configuration) configured to use SNMP bulk requests?

Markku

Yes its configured!and return data with snmp bulk
i dont have problem on snmp requests

[Markku]

Note that in Zabbix 7 there is the new walk[] item in SNMP:

https://www.zabbix.com/documentation/current/en/manual/config/items/itemtypes/snmp

That should ensure that all the SNMP metrics are collected at the same time.

See the "Mikrotik by SNMP" template for an example: there are the master walk[] items like "net.if.walk", and then the discovery rules use them to create the items, and they all get their data from the master item.

Markku

[sasanaalem]

Thank you but it doesnt help me!

[Markku]

Do you want to be more specific about your configuration and the effect of it when it comes to timestamps of the dependent items?

Markku