New Continuous Support System

An anonymous reader writes "eWeek is reporting on a new continuous open-source support system that helps to keep tabs on your mission-critical applications by providing constant diagnostic monitoring. The system is designed to match specific 'signatures' from your applications to a database of over 200,000 possible 'problem' signatures and alert the user for correction or analysis. From the article: 'SourceLabs' Continuous Support System features what Sebastian calls "adaptive diagnostic probes" that are fully integrated and configured for customer environments. The probes identify production issues and begin to gather diagnostic information to help get to the root of the problem, he said. Indeed, the probes can be configured so that as soon as a problem occurs, the SourceLabs support team extracts system information to find and resolve the problem. And the system includes a database of more than 200,000 signatures of problems that might occur.'"

Comments

Please Clarify

[Anonymous Coward]

I dont understand. Is this an advertisement?

Re:Please Clarify

[eln]

It's a fluff piece written by an "analyst" for a general-audience tech magazine, so basically it's a press release. If you look at other articles written by this guy, you'll notice that he is particularly fond of writing this type of "regurgitate the marketing" article.

Real News or PR?!!

[Anonymous Coward]

Was that a news article or a fluff press release? It'd be nice if the editors could let us know in advance when a slashvertisement plug is posted to the front page.

Lowering firewall

[Anonymous Coward]

Captain, I am receiving unusual data from the alien probe.

Analysis Spock?

Insufficient data. It may be a successful penetration from the Romulan sector. Or...

Or?

Or accounting is performing their end of month reconciliation jobs.

Puzzled

[Colin Smith]

Is it just me or is the FA completely devoid of useful information about exactly what and how the "SourceLabs Continuous Support System, technology " works? A non article. I have no idea how it differs from say Zabbix or Nagios.

Re:Puzzled

[bcat24]

Indeed. You can find a little more information on their website. Because putting a link to the company in the article summary would just make things too easy for people, right?

Re:Puzzled

[Bruce Perens]

When your production Java program breaks, it tells you, and Sourcelabs. Various sorts of breakage are detected. Generally the interesting problems are in the Open Source stacks that Sourcelabs supports, not in your own code, although the system can sometimes tell you when you are tripping over a well-known sort of error or an API calling mistake in your own code. Depending on the problem, you get an automatic message and/or you hear from your support person at Sourcelabs. Sourcelabe may give you a patch, advice, etc.

One interesting point is that you don't call customer service. They call you.Bruce

Re:Puzzled

[SanityInAnarchy]

So tell me again how that's different than, say, Nagios?

Insanely configurable -- can catch all sorts of problems. Can run a definable shell script when something breaks -- I'm not talking about "automatic message" or emailing someone at Sourcelabs, we had the thing configured to send an email/SMS to the main admin's phone. Cuts out the middleman -- the program calls me, I fix the problem. Works well when your "customer support" is often in-house.

Re:Puzzled

[Bruce Perens]

Nagios detects failures elsewhere. This instruments the insides of your Java program and tells you about many different kinds of failures that can happen in there, and it generally also tells you how to solve the problem.

I've seen this.

[MyLongNickName]

he system is designed to match specific 'signatures' from your applications to a database of over 200,000 possible 'problem' signatures and alert the user for correction or analysis.

The interesting thing is that no matter which 'signature' is noticed, the alert always reads "omfg n00b! read the fvcking manual!"

Wonder if there is a signature for

[fatboy]

http referrers from slashdot.org :)

Splunk with a different name?

[IGotYourSidekick]

How is this different from splunk? Now if it fixed problems for me...

Software is free, support is not

[DuckWizard]

I am really fascinated by this trend of selling support for open-source software. If a company creates a free, open-source product, and then uses support as their business model (RedHat, for example), doesn't that produce a conflict of interest in regards to the quality of their product? If the product is difficult to use, they will make more money off support. If it's rock-solid and completely intuitive, their revenues will crumble. Am I making any sense?

Re:Software is free, support is not

[bcat24]

I think they rely on two things:
1) Software almost always sucks to some degree
2) People are excellent at finding new ways to break "rock-solid" software

You know, the whole "make it idiot proof and someone will make a better idiot" type thing.

Re:Software is free, support is not

[mrbooze]

If the product is too difficult to use, no one will use it, and their revenues will crumble as well.

I think there is ample evidence in the enterprise software industry to contradict this theory.

Re:Software is free, support is not

[Anonymous Coward]

This only makes sense if you are talking about companies that sell support via time and materials, e.g. when something goes wrong, they charge you $XXX per hour.

In the case that you sell subscriptions, the software needs to be rock solid because you lose money on hard support calls. Large companies will still want support because of the case that something does go wrong, they need someone to call. Long term, as the market matures, you would expect support contracts to take into account the statistical chance of a errors in a system, and come up with appropriate pricing (similar to insurance).

The big news here, is that at SourceLabs we looked at the trend of supporting Open Source software and decided that long term this means innovating on support itself. If what you are buying is support, than your support companies should be competing with each other on how well they can provide this. This first set of tools and services is our way of jumping into this fray.

We do this by monitoring activities in the community such as new bugs, issues that come up on mailing lists, etc. We then index these, and digest these through a variety of automatic and manual methods. From this repository, we are able to both preventively warn customers of new issues found in both our internal certification efforts as well as new issues that show up from the community. In addition, we have probe technology that is able to pull information out of running systems when problems come up. One of the actions these probes can take is to match the information it pulled from the running system against our repsitory to see if the issue looks similar to one we've seen before or one that has come up on lists before.

As Open Source gets more traction and more companies come in to support it, we expect that there will be more competition over offering the best support. This will be a good thing, because it will force companies to fight over offering better services. As one of those companies, if you are able to be much better at
        1) Understanding the Risks of the software you are supporting
        2) Reducing the cost/time to resolution of an issue

you can begin to do a much better job of providing support than your competitors. This is our strategy. We want to make better technologies for providing support, so that we can be more responsive and can resolve problems faster. Long term, this will make support cheaper and more efficient. It will also make customers happier.

        Will Pugh
        Chief Architect, Sourcelabs

Huh?

[pongo000]

What kind of signatures? What kind of diagnostics? What the hell, exactly, is this article about?

And no, I'm not going to RTFA...if the submitter isn't articulate enough to succinctly describe what it is he or she is submitting, I'm not going to waste my time following the link.

Instead, I'm going to waste my time writing inane comments such as this...

Re:Huh?

[bcat24]

Don't worry, the article is almost as bad as the summary. You didn't miss much by not RTFAing.

Yea, right.

[BigCheese]

I can't count how many times I've heard this before. You either get spammed silly by alerts or turn the alerts down and then do what you did before you bought the product.

Sometimes you can get some use out of them but you've got to spend a whole lot of time with it in setup and ongoing adjustments.

Too many managers buy these things expecting a "Magic Bullet" solution.

Re:Signatures?

[Bruce Perens]

And, if you have something like this, you probably don't want the problem to occur at all.

You may not be the first customer to hit the problem. Also, the problem can manifest itself in a non-signature-dependent manner, like throwing an exception. Then if you are not the first to see it, signatures may come in to play in telling you why the exception happened.

Zenprise

[sfcat]

I previous worked for a company where I developed something very similar to this 'Continuous Support System'. But it was targeted at Exchange (MS, boo hiss, I know, I dislike them too).

Anyway, it was a very interesting and difficult problem. One of the biggest rubs was the level of assurance you had to provide. In otherwords, can you let the system make changes on its own or should it just recommend changes? If the system mis-diagnoses even one problem, it might break more stuff than it fixes. Most monitoring tools have big problems with 'false positives'. Add to that that the system can't necessary 'undo' all changes. Our solution was to allow the administrator to run the system in a variety of modes so they could choose if the system applied the fix automatically, with approval, or just suggested how to fix the problem.

As for how the system actually works, it basically takes a middle approach between ML (machine learning) and KR (knowledge representation). Basically, either you can hard code all the types of problems you have in a KR language, or setup some big neural net (or other ML algorithm) and let the system 'learn' problems. We split the difference and added some domain knowledge. Certain types of 'features' (parts of a diagnose such as the disk is slow) were diagnosed by ML algorithms, but ultimately KR rules written by Exchange experts actually diagnosed the problems and suggested repairs. A very time consuming, but more reliable solution (but less cool).

Re:And the most common problem is...

[NumbThumb]

It's always PEBCAK. The tricky part is finding out whose keyboard.

The Phone Conversation

[Jerky McNaughty]

Here's the one-sided phone conversation, as heard from a neighbor of the support person at SourceLabs.

Hey, is Arnold around? This is Frank over at SourceLabs.
Hey, Arnold. It's me again. How's it going tonight?
Oh, really, it's 2:30am there? Wow.
Yeah, I know what you mean. Yeah, it's raining here in Seattle, of course.
Hey, listen, the reason I'm calling is because your shit, yeah, yeah, it's crashing again.
Hey, don't blame me. Talk to your manager about it.
Well, he's the one that bought this support.
Listen, though... the stack trace pops up on my screen here and I'm supposed to give you a call.
Well, yeah.
Yeah.
I mean, it's 24x7. You're somewhere in that 24 and somewhere in that 7, so here I am.
Yeah, I don't enjoy this either.
I know what you mean.
Well, the stack trace looks like your Oracle database is hosed again.
Yeah, tell me about it.
Well, you're using the thin-client drivers.
Looks like you can't get any JDBC connections. What a bitch.
I know, sucks that your site is down. What a pisser.
Well, most people monitor this kind of basic stuff on their own.
Yeah.
Uh huh.
Well, maybe some log4j and Nagios would work. Or something.
Yeah, really. It'd save the time it takes me to call you. Good thing you're only taking like 100 orders/minute at this time of day. Heh heh heh.
Yeah, I had to wake my ass up early this morning, too. I'd almost rather be doing drywall at the new McDonald's.
Yeah, ok, cool. Well, see if you can get your Oracle P.O.S. back up again.
Definitely.
Cool.
Well, I'll probably talk to you soon. Bye!