Our new Transaction Trace Warehouse ("TTW") for AppInternals Xpert is striking a chord with DevOps teams. But I didn’t realize until recently the magnitude of impact it has on shrinking software release cycles. Nor how broadly accessible it would be for large audiences within an IT organization.
I spoke to over
100 people who saw demos at OPNETWORK 2012 in Washington DC
and VMworld 2012 in San Francisco within the last month. Their feedback helped
me understand what our early adopters of TTW told me before: "Wow- this fundamentally changes how operations and
development staff work together.” So what is this big idea? Here is a summary from my notes, below...
Developers tend to be busy, and they communicate with QA and Operations in iterative steps: new features get rolled out frequently to meet new business needs. Sometimes they are tested well in QA, sometimes not. When performance problems occur in production, a finger-pointing game ensues between IT divisions. Once initial troubleshooting is complete, developers need to verify the problem and use traditional software debugging and testing tools to fix things. And then the cycle begins again. Our customers complain that these cycles can take months of tedious work, with new releases often coming out before old performance issues are solved. If a third party software company or contractor is involved in the release cycle, even more fun! A shorter cycle means better performance and more ROI from the application. TTW for AppInternals Xpert dramatically improves this dynamic by providing any authorized staff from Operations or Development/QA with hard data they can turn easily into action.
Here are a few key enabling capabilities:
Developers tend to be busy, and they communicate with QA and Operations in iterative steps: new features get rolled out frequently to meet new business needs. Sometimes they are tested well in QA, sometimes not. When performance problems occur in production, a finger-pointing game ensues between IT divisions. Once initial troubleshooting is complete, developers need to verify the problem and use traditional software debugging and testing tools to fix things. And then the cycle begins again. Our customers complain that these cycles can take months of tedious work, with new releases often coming out before old performance issues are solved. If a third party software company or contractor is involved in the release cycle, even more fun! A shorter cycle means better performance and more ROI from the application. TTW for AppInternals Xpert dramatically improves this dynamic by providing any authorized staff from Operations or Development/QA with hard data they can turn easily into action.
Here are a few key enabling capabilities:
- Let each team see any transaction any time they want to in production -- this comes from our Big Data approach to capturing, storing, and indexing every real user transactions (instead of just sampling).
- Make the transaction search capability so intuitive that users can get started immediately after being exposed to the product, without any formal training.
- Deliver code-level detail and performance analysis for each transaction as it was executed across different physical and virtual systems. This way, developers can act immediately to make changes in the right places.
- Automatically detect anomalous behavior that affects performance in production, so finding the right transaction to examine with TTW can happen proactively.
One of our power users was
recently promoted to head up the web architecture function for his company's multi-billion
dollar customer-facing website after he used AppInternals Xpert for years to holistically
analyze the site’s performance. The brand new TTW capabilities prompted him
to expand the deployment of our AppInternals agents from "many
dozens" of systems to "every application component that touches the
website.” Another power user told me, “we published a link to TTW on our
intranet; over 400 people access it regularly to see live transactions; it’s
self-service now, meaning I don’t have to be in the middle of every
troubleshooting exercise.”
I think we’re on to
something here…
Posts
No comments:
Post a Comment