Dan, consider lighting up a websvn instance in parallel to obs, pointed at your rancid depot. it’s got a nice layout, and you get a bit more in the granularity toolbag (e.g. permissions). our auditors love it. you lose a bit of the SPoG effect, but really it’s an ideal drill-down.
aside:
top 3 wish list (e.g., after obs alerting gets finished):
- eigrp feature-parity with ospf/bgp - f5, palo & checkpoint device support (rancid is weak here also, so obs is in good company /grin/ ) - auto-thresholding sexiness (a math-inclined list member offered to contribute some framework for this a while back?)
alerting wishlist items:
- bark at me when the route count of established routing peer sessions (esp. bgp) changes from $baseline - bark at me for threshold deviations, esp. unnaturally lower-bound ones (e.g., ckt that normally pumps nMbps flatlines to zero) (tells me my edge peers probably did something stupid that they will blame me for anyway. knowing about it first helps. LOL).
ps:
- scale of syslog ingestion (or possibly the regurgitation thereof) might warrant some love. 7MM/day entries in volume makes for dreadfully long page loads.
- anyone have the mysql agent script poking at an instance of mariadb? are your results in obs fully populated? (processList, innoDBinsertBuffer, innoDBsemaphore,innoDBxactns?)
- am working on some integration ideas with gestio & rwhois, so that subnet links in obs, esp. for rfc1918 space can become useful info (has anyone gone down this road already?)
cheers,
R.
On Jan 29, 2014, at 21:34, observium-request@observium.org wrote:
From: Daniel Rolfe daniel.rolfe.au@gmail.com
Is there any way to give config access to global read accounts ?