Graphs for same interface on v0.8.8a & v1.1.38 are different
Moderators: Developers, Moderators
Graphs for same interface on v0.8.8a & v1.1.38 are different
I have two different version of cacti, v0.8.8a and v1.1.38. And I found that the Traffic Graph behave different. It seems the peak values for v1.1.38 are smaller that old cacti version. The old cacti have been used for few years and the its values seem correct. Could any expert can help me what's wrong with the new cacti?
- Attachments
-
- v1.1.38_2.png (145.41 KiB) Viewed 507 times
-
- v0.8.8a_2.png (115.38 KiB) Viewed 507 times
Re: Graphs for same interface on v0.8.8a & v1.1.38 are diffe
Hi sat38, i´m no expert but i can think in 2 possibilities:
1-the 245 MB is a false reading (spike to kill). It casualy hapend to my yesterday with a peak of 80GB in an interface that can´t exceed 2MB.
2-even when your are monitoring the same interface there are 2 diferent polling cycles that probably are not in sync between each other, so the new cacti poller could miss the 245MB peak but the old cacti get it.
1-the 245 MB is a false reading (spike to kill). It casualy hapend to my yesterday with a peak of 80GB in an interface that can´t exceed 2MB.
2-even when your are monitoring the same interface there are 2 diferent polling cycles that probably are not in sync between each other, so the new cacti poller could miss the 245MB peak but the old cacti get it.
Beast
GitHub: DoubleBeast
GitHub: DoubleBeast
Re: Graphs for same interface on v0.8.8a & v1.1.38 are diffe
Hi Beast, I have another SNMP server to monitor the traffic, the traffic value from it is very similar to v0.8.8a, and the peak value is actual correct. I guess the graph from v1.1.38 does not show the actual peak. Since I would like to know the internet bandwidth is sufficient to handle the peak traffic, it is important that the graph could show the actual value. Just don't why on v1.1.38 has so much different. Is it a bug or configuration issue?
Re: Graphs for same interface on v0.8.8a & v1.1.38 are diffe
-verify if the new cacti have the graphs set for 64bit counters (should use those when the traffic is above 100M)
- i don´t know if this is an option for that interface but if you can try to force the traffic for a couple of polling cicles (or simply use realtime) to check if it register the higher bandwide usage.
- i don´t know if this is an option for that interface but if you can try to force the traffic for a couple of polling cicles (or simply use realtime) to check if it register the higher bandwide usage.
Beast
GitHub: DoubleBeast
GitHub: DoubleBeast
Re: Graphs for same interface on v0.8.8a & v1.1.38 are diffe
Hi Beast, I realized the graph has 32-bit & 64-bit different after I read your post. And I now removed all previous graph and set the new graphs to 64-bit, will see how will be the values in the next couple days. Thanks for you help anyway. Cheers!
Re: Graphs for same interface on v0.8.8a & v1.1.38 are diffe
Hi Beast, the problem was fixed by using 64-bit graph, since the old version doesn't have this 64-bit selection, I think this is the part I missed. Thanks for your help!
Re: Graphs for same interface on v0.8.8a & v1.1.38 are diffe
Excellent sat38! glad i could help you.
Beast
GitHub: DoubleBeast
GitHub: DoubleBeast
Who is online
Users browsing this forum: No registered users and 0 guests