My company has been running Observium for the last 5 years or so to monitor our core and edge network, plus managed customer devices, and this includes our upstream peering links (we're a small ISP). We occasionally get tiny outages reported by some customers, where they might lose connectivity for 30-60 seconds. Unfortunately, the customers might only be doing 50-100Mbps at the time, and we're normally pushing 3Gbps over our main peering link. When you combine that with Observium’s 5 minute polling interval it means these "outages" are impossible to see on the core links.
I've seen it's possible to tune Observium to a lower polling interval, but that affects every sensor, and we're monitoring a lot of stuff so the load on the server would increase massively. The only other NMS I've used extensively is PRTG but that's outside of my company’s budget for the time being, but that did at least allow you to set custom polling intervals on individual sensors.
So, my question is, what are people’s recommendations for network monitoring? Windows or Linux based, either is fine. It doesn't have to be free either, there is some budget for this. It'll be monitoring mainly Juniper but also some Cisco and Extreme, around 100-125 devices total.
Thanks in advance!
No comments:
Post a Comment