Aggrgation, Nagios Check, Graph Caching
Hi,
After using Observium for some time I have some ideas and questions:
* Is it possible to create aggregated graphs? F.e. I'd like to aggregate all uplinks (which are on different routers) into one big graph (all uplinks as a single line, in one graph (overlay). And the traffic aggregated).
* This could maybe end up in a Dashboard, which shows me an overview of what's going on in the network. (Aggregated Graphs, Top Ports, Recent Syslog, High Values, ...)
* I'd like to check some values collected by Observium with Nagios to generate an Alert (F.e. the temperatures, dBm, ...). This could be a plugin like "check_observium <device> <type> <thresholds>". How can I access the raw data in Observium from outside? Is there something like this available, or does it need to be written? My main concern is to have just one tool which Polls the devices and get the collected data out of it. I don't want to have many tools which are all doing SNMP. One single tool should be enough.
* How can I get the graphs out of Observium to display in another tool?
* Are there any caching mechanisms for the Graphs? Everytime I hover over a link which creates a Pop-Up the graphs are being generated. What about caching the generated graphs for f.e. 5 minutes?
Cheers and thanks a lot for this great tool!
Tobias
On 31/12/2012 03:44, Tobias Brunner wrote:
Hi,
After using Observium for some time I have some ideas and questions:
- Is it possible to create aggregated graphs? F.e. I'd like to
aggregate all uplinks (which are on different routers) into one big graph (all uplinks as a single line, in one graph (overlay). And the traffic aggregated).
We have a multi_bits graphtype which allows you to do this. You send it a comma-separated list of port_ids, and it draws a big graph. It's the same graph used for the device_bits graphtype.
- This could maybe end up in a Dashboard, which shows me an overview
of what's going on in the network. (Aggregated Graphs, Top Ports, Recent Syslog, High Values, ...)
Less useful than you think :)
- I'd like to check some values collected by Observium with Nagios to
generate an Alert (F.e. the temperatures, dBm, ...). This could be a plugin like "check_observium <device> <type> <thresholds>". How can I access the raw data in Observium from outside? Is there something like this available, or does it need to be written? My main concern is to have just one tool which Polls the devices and get the collected data out of it. I don't want to have many tools which are all doing SNMP. One single tool should be enough.
We don't currently have any method of doing this, but it wouldn't be difficult just to kick these values out with a bit of php/perl. They're just numbers in MySQL. We store counters and /sec values for bits/packets/errors for ports, and the measurements of most other entities in the database.
- How can I get the graphs out of Observium to display in another tool?
You probably want to proxy them with a php script. You can configure observium to allow unauthenticated graph access from a host/network allowing you to proxy from another webserver.
- Are there any caching mechanisms for the Graphs? Everytime I hover
over a link which creates a Pop-Up the graphs are being generated. What about caching the generated graphs for f.e. 5 minutes?
Nope. Far more hassle than it's worth, believe me.
adam.
participants (2)
-
Adam Armstrong -
Tobias Brunner