In light of the recent major outages experienced by XO and Abovenet, I
have become very interested in finding a 3rd party vendor to monitor the
availability of my network from a variety of locations.
I've looked at keynote / redalert, but oddly enough they don't seem to
be able to monitor simple ICMP round-trip times - they (or at least the
sales folk that I've spoken with) insist on checking the status of a TCP
port or the response time of a web server.
Google searches have yielded a few leads, but nothing terribly
promising.
Can anybody recommend a vendor for this type of service? I suppose I
could always just get a bunch of VPS accounts here and there and run
smokeping, but I'd really like to avoid that.
Check out WebSitePulse.com While they too are geared for service testing (such as HTTP, DNS, IMAP, etc); they do offer a barebones ICMP check (with support to pull from multiple geographical locations) with some basic graphing & escalation support (looks like they can notify when a predefined threshold of round-trip time is exceeded).
Keynote should be able to monitor simple icmp echo requests. Last time i
worked with their product, i was able either to monitor our network with
icmp echo requests or measure the performance of our web-farm with
special URLs.
Anyhow, as i grabbed for any provider for this case, keynote seems to be
the only good one with reliable values.
We've tried repeatedly to get a more advanced setup with Keynote, however in the past 8 months we've yet to be able to get anyone to reliably contact or respond to us. Horrible service, and I guess new business doesn't matter to them. When we did get a hold of someone, they mentioned they could support simple ICMP requests.
Can anybody recommend a vendor for this type of service? I suppose I
could always just get a bunch of VPS accounts here and there and run
smokeping, but I'd really like to avoid that.
Numerous times their alerts have beat out our notifications without false alarms.
Also, as the reason one purchases such a service, its a backup for our own
monitoring. We also use Keynote as thats what our customers are used to, and
also subscribe to. This allows us to se what they see should we need to discuss
SLA's. Systems Watch is much cheaper than Keynote for those that have actually
priced out, or subscribe to the service.
All in all, when I get an alert from Systems Watch, I do make way to a console
more than I would with my own monitoring.
Regarding Keynote support - the company I work for pulls alerts from the RSS feed they offer. They changed the format of the feed a couple of months ago without notice, breaking our monitoring and leaving us blind for some time. When we contacted our service rep, he wasn't even aware of the changes they made. Your mileage may vary.
To them "simple" means it's just a ping check. They won't montior/graph/care about latency.
I was pondering creating a "smoke ping collective". Get a bunch of guys to agree to run smokeping and monitor each other. That's a great tool for visualizing changes in latency and works just as well with ICMP as with HTTP.
When we did get a hold of someone, they mentioned they could support simple ICMP requests.
To them "simple" means it's just a ping check. They won't montior/graph/care about latency.
I was pondering creating a "smoke ping collective". Get a bunch of guys to agree to run smokeping and monitor each other. That's a great tool for visualizing changes in latency and works just as well with ICMP as with HTTP.
There is this really awesome project from RIPE (like usual
Please check, and start using RIPE TTM: http://www.ripe.net/ttm/
See the site for presentations, tools, info, etc etc etc etc...
One app I like a lot is Ping Plotter, but it only runs on Windows, so it isn't good for remote monitoring. We do use it for some things, however. I like the detailed traceroute / latency visualization it has. It also has a hard time with a lot (100+) nodes being monitored. SmokePing works well, but lacks detail in the form of traceroute.
My bad, you might be able to do it with PingPlotter using remote proxies that are linux. I can see using the Vixie personal colo list to find cheap vm offerings in various locations. Other option, a few could get together and share some resources to get the proxies distributed.
My bad, you might be able to do it with PingPlotter using remote proxies that are linux. I can see using the Vixie personal colo list to find cheap vm offerings in various locations. Other option, a few could get together and share some resources to get the proxies distributed.
Did you actually *check* the URL I passed in? TTM does quite a bit more and is already distributed around the world and available to ISP's.
Again:
There is this really awesome project from RIPE (like usual
Please check, and start using RIPE TTM: http://www.ripe.net/ttm/
See the site for presentations, tools, info, etc etc etc etc...
I did look at it, it still lacks a few things, but it does cover most. It would be nice if you added some screenshots or demo pages as to what the reporting looks like. I had to dig around and find a paper on the slammer worm to see what the output looks like.
I did look at it, it still lacks a few things, but it does cover most. It would be nice if you added some screenshots or demo pages as to what the reporting looks like. I had to dig around and find a paper on the slammer worm to see what the output looks like.
Contact RIPE NCC for that, if you would have read it, you would have found that it is their project.
