Best Web site monitoring service?

by philg

Folks: We’re investing in a major hardware/software upgrade at photo.net. It would be nice to know if we’re doing better or worse overall when it is done. What are the best external Web site monitoring services these days? I want to know what our average time to serve a page is before and after the upgrade. Would also be nice if Jin’s phone could be called when the site is unreachable.

19 thoughts on “Best Web site monitoring service?”

  1. A friend of mine used to work for Witbe, and he recommended their services to me. I have never used them though. “Witbe on demand” costs 90 euros/month, and monitors the site from various locations.

    http://www.witbe.net/

    Florian.

  4. Phil, why all the questions about the photo.net server? I thought you were no longer involved in the community? Have you bought it back?

  6. Phil, As a publisher, what is your perspective on the impact of browsers (Firefox) that effectively block all ads? How does the presence of ad-blocking technology impact you as a publisher? Does it lead you to consider serving up different experiences to users of different browsers to somehow disincentivise ad-blocking? Does it lead you to consider technical counter-countermeasures (whatever these may be)? You’ve had significant experience with most of the many roles involved in bringing content to end users on the web, so I’d be interested in hearing what if any thoughts you have along these lines.

  7. There is http://www.fiveruns.com, a systems monitoring service with nice UI and a ‘web 2.0’ feel to it. You can sign up for a 30-day trial. (annoyingly, no prices published on the site)

    Please let us know which service you decide on ultimately and why. Thanks!

  8. Phil .. I recommend you rent a 128MB RAM, VPS (virtual private server) for $25 or so at a reliable hosting company and run your own monitor tool (maybe the uptime package).

    If you want a recommendation for a simple, lightweight, monitoring package that can both send email and do pages (as well as escalation), I suggest Argus, which despite being written in Perl is fast and for what it does, lightweight as it fits in about 20MB as a single process which does SSL, normal http, DNS, SNMP etc. monitoring.

    Argus URL: http://argus.tcp4me.com/

  9. McDavis: I run Firefox and I see ads on photo.net. I don’t see this as a big factor. We are doing ourselves a lot more damage by not having reelvant/interesting ads than Firefox is by suppressing some ads. Everything on a page served by photo.net should be relevant and interesting to the reader, including the ads. If the ads are relevant and interesting, people won’t want to block them.

    Patrick: I already have a pizza box at a different ISP to run http://www.greenspun.com. Does this Argus software gather and present the statistics the way the commercial packages do? I’d just like to see average response time, I guess. If we run our own installation, we could do nagios, yes? Or is that more for the guts of the operating system?

  10. Get the free one-month Keynote trial 15 days before upgrading…

    They are the gold standard for monitoring reachability from multiple backbones. I wouldn’t recommend running your own performance monitoring, but you should run service availability monitoring from two different locations. Nagios is currently the open-source monitoring platform with the most traction.

  11. seems like most serious shops I’ve seen use nagios. it’s sensibly designed, if a touch complex. last I checked it was missing built in rrd/graphing/trending, but it has a facility to integrate third party trending (termed “perfdata” or “performance data”), and there are some off-the-shelf packages designed around it… last I checked they seemed not quite mature. I ended up somewhat hacking one called “nagiosgraph”, and only to collect the data, not graph.

    for graphing I use drraw, which imo outshines most other trending packages mostle due to it’s regex enabled configuration model… a big win for a system with more than a couple nodes. drraw also has a web-ui graph configurator, but stores it’s config in flat files which are readily version controllable… also a big win imo.

  12. Montastic: the free website monitoring service.

    Website monitoring made cool:
    # Get an email when your site goes down
    # Get an email when it goes back up
    # Read statuses via RSS or Yahoo widget
    # No unreasonable limit on the number of websites monitored

  14. Personally, I’ve never been happy with 3rd party services for reasons too many to list here. Once I came to the conclusion that I was done with 3rd party services, I did some hard searching for software and applicances to do the job for me and ultimately decided on a company called Science Logic about a year ago. http://www.sciencelogic.com/

    We have three devices strategically placed, two in the US and one in EU.

  15. I’ve had good luck with both gomez ( http://www.gomez.com/ ) and keynote ( http://www.keynote.com/ ).

    I think the user interface on gomez is a little nicer, allowing you to profile the response time of individual page components. Gomez at one point had a link with Empirix, another service worth checking out.

    All of these guys have been around for years, and are used by Really Big Companies.

  18. As a long time follower of acs and photo.net, I would like you to try the server uptime monitor service at http://basicstate.com/

    It reports response times, but you must bear in mind that this is a relative measure as opposed to an absolute measure vis-a-vis the response times to clients using differing connection options.

    Both primary and escalation alerts can be sent to any phone as long as you have a SMS email address available to plug into your control panel. I would recommend that the SMS only be tied to the escalation alerts.

    Since the service is free, there’s not much I can do to make it a special offer just for photo.net, but thanks for all the good reading in the past 🙂

Comments are closed.