It’s been many many months now since Quest Software picked up Magnum Technologies. Where’s the beef? Where’s the new Business Service Management (BSM) story? How’s the Magnum Technologies solution integrating into the Foglight 5 architecture? Are there significant architectural challenges preventing this?
Like I mentioned last year, I can tell that you have added some of the things I taught your PM’s about way back in 2004-2005, but where’s the new story? Where’s Foglight 5.x/6? When will you bring a new BSM story to the market? This is a bit stale and doesn’t look like the Magnum Technologies stuff is incorporated.
I realize there may have been challenges with the NASDAQ distractions and M&A activity in other areas of your company. Maybe you might consider donating your Foglight and/or Big Brother suites to the Open Management Consortium (OMC) for oversight and development? They’re in search of a good Open Agent and I think you’ve got a great platform and framework for just that.
I must say though that the Big Brother site has a very nice look and feel! Unfortunately, the product as demo’d on your site is still not fit for executive consumption.
Comments on this entry are closed.
I could not agree with you more. My company evaluated the Foglight 5 solution and as where it stands today, it has serious scalability issues and lot more can be done on process flows of how solution identifies performance issues. Generally speaking, I believe it has a long way to go before it gets accepted by market.
Hi Doug, I enjoy your column and must apologize for not keeping you up-to-date with Foglight. We released Foglight 5.0 in March of 2007 which included an architectural rework to make Foglight 5 a model-based management system. This has allowed Foglight to handle dynamic and complex environments with less administration and manual work than normal, and has also made the job of integrating other management systems quicker and easier. In mid-2007, the Magnum products were integrated with both Foglight 4 (so our existing customers could get the benefit) and Foglight 5 (for new prospects and customers) and are being used by many of our customers.
John – I’m very sorry to hear that you had a bad experience with Foglight. We did have some scalability issues in early releases with certain types of environments and situations. However, the monthly release schedule that we’d instituted with the release of Foglight 5 paid off here as it allowed us to quickly address those issues. Customers using our most recent Foglight 5 versions have been impressed with its scalability, stability and performance. In fact, we’re in the midst of implementing Foglight 5 in a >3000 CPU datacenter for a company that was looking for a single pane of glass view of their user experience and how it was connected to application, database and infrastructure performance. I’d love to chat with you (brad.micklea@quest.com) about the problems you had with Foglight and get your honest feedback on where we can and should improve – we’re determined to be the best critical applications management solution in the market but we’re not going to get there without taking both the good and bad feedback to heart.
On the plus side customers who’ve been working with Foglight 5 have told us that we have some standout capabilities including:
– Our comprehensive end user management capabilities which combine synthetic transaction with real user information in a single view
– The correlation we offer between those user views and the applications, databases and infrastructure that support the services they use
– The ease with which customers can get customized, role-specific dashboards from high-level BSM views, to deep root-cause diagnosis views in one place
– Our ability to quickly incorporate data from custom scripts and other applications being used for management in the IT environment
Network Computing recently did a non-paid and self-directed review of several application management solutions where they said “Quest’s offering is a bright spot in our search for truly holistic application management.” (full review at: http://www.networkcomputing.com/channels/appinfrastructure/showArticle.jhtml?articleID=204300081).
Admittedly we’ve been keeping the Foglight 5 fanfare to a minimum but we’re planning a big launch very soon. 2007 was Foglight’s biggest year in terms of revenue and number of new customers and 2008 is starting off equally strong. Stay tuned.
Brad Micklea
Manager, Foglight PM
I couldn’t disagree with you more, John.t – I’ve been a Foglight user since version 3 and have recently implemented version 5.1.6. The latest version scales well and the integration with FxM and FxV is exceptional. The Magnum components are currently being integrated into Fog5.
Wow, great dialog guys! Keep it up, keep sharing, show us what’s improving, tell us the scale per FGL server (FGL 4 couldn’t handle more than 200 servers per central FGL server), show me the BSM stuff!!!!!
Doug
Foglight 5 has surpassed the limitations of 4 – the actual quantity of servers is determined based on the number and type of agents being run on the host. We do a sizing exercise with prospects to ensure that they’re not going to see scalability issues but typically we’re seeing >1000 linux servers or >350 Oracle DB instances with a single mid-range management server. We’re also introducing a server federation feature in our next release (later this quarter) that will allow you to get a unified view across multiple Foglight management servers. This will allow people to see a unified view of their entire IT environments (dashboard, reporting, alerting, etc…) regardless of the number of management servers used underneath.
Cool, sounds like things are improving. That federated layer was desperately needed!! It was IMPOSSIBLE to manage the 20+ FGL4 servers. Fortunately, we built our killer provisioning and management tool and custom Netcool probes to make things somewhat manageable. I’ll assume your CMDB component plays a role somehow in relation to the FGL and Magnum stuff. Did you ever decide on releasing the Agent SDK?
Doug
Brad, thanks for sharing these details. Since you have asked , last i saw that Foglight 5 Java and Oracle agents were polling somewhere around 5 min. Which means i had to wait 5 min or sometime 10 min to get some alerts popped up. When i lowered the sampling frequency to 60 seconds, thats where i s saw a significant overhead on my Weblogic servers and Foglight server database grew @ 10MB data space / day and Foglight database could not keep up by 2 Java agents and 3 Oracle agents sending data @ 60 seconds frequency . In the latest build of Foglight, is the frequency still 5 min ? Secondly when will PerformaSure be part of Foglight 5 ? I was told that PerformaSure agents and Foglight 5 agents cannot live together ? Whats that status on that
Thanks for the clarification John. The agent polling is configurable as you saw but the default remains 5 minutes. Within Foglight 5 it’s also possible to change the polling for specific agents or environments so that you can tailor your data collection to your needs. As I said the scalability of the latest release is much improved and we’re seeing hundreds of agents without incident. We also revamped the data management layer of the Foglight server to use periodic roll-up and data archiving. As a result you’ll certainly see high database growth for the first week or so but at that point the roll-ups and archiving will be in full swing and the database size will flatten out.
On the PerformaSure front the Foglight 5 agent for Java was completely rewritten and now has all the capabilities of PerformaSure built-in. The only difference between the Foglight and PerformaSure agents is that the Foglight agent was tuned for production monitoring and so gathers transaction traces less frequently and in a more focused fashion to keep overhead low; the PerformaSure agent has been tailored to pre-production environments where its more frequent trace gathering is needed. The majority of production-focused use cases we saw our customers turning to PerforamSure for are now fully available in Foglight 5 with the added benefit that the traces themselves can be queried side-by-side with metrics data to create dashboards that included aggregated and single-user trace data side-by-side.
Doug,
The reason for Quest’s latency to roll out a BSM story is directly related to ineptness in the Product Marketing organization, namingly the under management of such group. Brad most likely cannot respond to this publicly, for obvious reasons, but rest assured until Quest addresses this marketing mis-direction, don’t expect Foglight ubiquity.
The FogLight Management server still has performance and scalable and I am running the 5.5.4 release. The amount of resources, (CPU and Memory), needed for running the FMS and DB has increased significantly since FogLight 4. Also, the truth behind the statement from brad, “On the PerformaSure front the Foglight 5 agent for Java was completely rewritten and now has all the capabilities of PerformaSure built-in. The only difference between the Foglight and PerformaSure agents is that the Foglight agent was tuned for production monitoring and so gathers transaction traces less frequently and in a more focused fashion to keep overhead low; the PerformaSure agent has been tailored to pre-production environments where its more frequent trace gathering is needed.” is not complete. PerformaSure tries to capture 100% or the metrics for 100% of the methods for 100% of the requests, while FogLight can trace only a single request. The agent may be the same, but you cannot execute all of the PerformaSure Use Cases using FogLight. Analysis of the information puts a big hit on the FogLight Management server instead of doing the analysis on the fat workstation. Also, when you are debugging an issue with PerformaSure, many time you gab small sessions and set the sample rate as low as possible and analyze the information off-line. With FogLight 4, we were able to automatically grab a 2-5 minute session based on any alert. having the complete trace within a minute of an alert was invaluable. And this was done for me while I was sleeping. With FogLight 5 we seem to try and catch the trace the next time it happens. Adding the federated FMS adds much more Infrastructure and support cost to the implementation.
I have used Foglight for 2 years and Foglight 5.5.4 does perform much better than previous versions of Foglight–but the performance is still embarrassingly poor and the software is still unreliable. Many of the Foglight products have been forced into the terrible WCF causing an untold number of issues that make the software painful to try to use, and probably worse, the captured information is many times misleading, missing or just wrong…at least with the products I’ve used. Buyer beware.
I guess the software does do some of what it promises despite it’s setbacks. In my opinion, it’s under no circumstances worth the effort. If it were up to me, I would remove this software from the entire implemented infrastructure in our environment.
Maybe Quest will fix these problems in the future since they have made a lot of progress. From my experience, I wouldn’t trust them though.
We evaluated Foglight ver 5 few weeks back and was not at Ll satisfied with it’s data collection metrics . It does have great capability of dashboarding but we determined you really need some serious development expertise to use the dashboard. They do have some template but templates does not help when you want to customize the products. It’s also lacks agents to cover tibco, and enterprise service bus monitoring. Our conclusion was that this software may be ideal for a small company as it does not scale up as compare to other tools we evaluated from HP and CA