Enhancements to SNMPCollector #63
Conversation
Modified the base SNMPCollector to do a lot of what the SNMPRawCollector does. This means that it can handle both individual GETs and WALKs via configuration. Added tests for all functionality in this class.
…t conversion for non-int types
Conflicts: src/collectors/snmp/snmp.py
|
Hi - This PR appears to have PEP8 issues. Would you mind cleaning that up? Check the Travis build log for more information. |
|
Absolutely. I'll take a gander at the failures. |
|
I'd like to revisit this PR @shaunduncan is this still something you're willing to clean up and help get merged in? |
|
@jaingaurav absolutely! I actually forgot I had this sitting around. I'll merge the upstream branch and re-push |
Conflicts: src/collectors/snmp/test/testsnmp.py src/collectors/snmpraw/test/testsnmpraw.py
|
@shaunduncan so sorry, mind rebasing this one more time? It looks fucking awesome, just wanted to fix a few snmp bugs before this dropped. Sorry! |
|
Sure does bring back old memories when I stumbled across this PR. @josegonzalez @shaunduncan If shaun doesn't have any interest in this project anymore, I'd be willing to pick up the PR and get it rebased and updated for the current state of the code. Shaun and I were working together at a previous employer when this was first opened, but I actually stumbled across it this morning when I was trying to add monitoring of my cable modem and router to a Diamond instance I run on my home computer. I was a bit dismayed by the current state of the SNMP Collector, so I'd be happy to get it working and update this if there's a chance of it being merged and released. |
|
@jantman Feel free to rebase this and open a new PR. Reference this one when doing so. |
In relation to Issue #59, I've taken the liberty to make some changes/enhancements based on some of the behavior I've seen. I couldn't tell if there was a major difference between the SNMPCollector and SNMPRawCollector, so I've combined that into just the base SNMPCollector. I've also included a large amount of tests with this as well. The changes here should be backwards compatible as some of the method signatures have not changed. I've also tested this on a VM running snmpd using both this improved SNMPCollector and the unmodified SNMPInterfaceCollector.
I've also added an enhancement to perform SNMP walks via a configuration. For example:
Would produce metrics like
foo.bar.1andfoo.bar.2.1.1.1.Please let me know if this change makes sense (or if it's just a terrible idea). I'd be happy to hear the feedback. Thanks!