Loupe 4.0.1 Released With Slack & HipChat Notifications

We’ve updated Loupe with a number of small improvements & defect fixes along with big improvements to notifications.   This upgrade is already running for our Loupe Service customers and is available for download to our self-hosted customers.

Enhanced Notifications

We’ve done three major improvements to notifications:

  1. Users can self-manage their notifications:  Any user can go into their user profile area and set up their own custom notifications without being a site administrator.  Administrators can still manage everyone’s notifications.
  2. Slack & HipChat notifications:  You can route event information right to Slack & HipChat for the whole team to see and react to.  More on this in our next blog post.
  3. Multiple targets for notifications:  Previously you had to specify a single target for a notification with the rule.  If you wanted multiple people to receive the same notification you either duplicated the rule or set up your own external distribution list.  Now you can have any number of users, email addresses, or services informed by a single rule.

Enhanced Performance

We’ve been migrating all of our systems to Azure and along the way have found the performance on SQL Azure hasn’t been as good as we expected compared to SQL Server.  This pushed us to spend the last two months revising the way we query data, indexes, and the schema itself to work better within the constraints of SQL Azure.  The net effect is that we’ve reduced disk IO requirements by up to 50% compared to 4.0.  If you’re running your own Loupe server you’ll see better performance and lower impact on your SQL Server.

Some customers will see dramatic improvements in performance - in particular where there are a very large number of unique events being tracked by the system.   For example, we’ve seen customers that have several million unique events (which is unusual and typically means the log messages are overly unique - a perfect use case for our Log Event Redaction feature!) and have never gone through the review cycle.  This caused very slow performance when determining the most important events for review and a few other key features because too many events were being considered by the database.   We’ve made changes so even in the most extreme cases these views should complete in at most a few seconds.

Many Small Improvements

I’d recommend you read the release notes to see the changes we’ve made to the web and Windows UI’s.  There are a number of changes that may seem small in the list but can really improve how effectively Loupe can help you solve problems.   For example, searching in the web UI now automatically selects the first tab with results so if you put in a computer name it’ll jump to that tab once it gets a result.  Our favorite improvement is the integration of server events into the session display in the Live Sessions view of Loupe Desktop.  Now when you click on a running session you can see what unique errors and warnings it’s recorded without opening it which is a very quick way to understand if a server’s running normally or not.

We’ve also packed more defect fixes into this release than anything else we’ve ever done.  These issues were mostly in the web UI and mostly related to multitenant issues in our SaaS or Enterprise Editions.  Loupe 4.0 had a completely rewritten, and substantial, web UI so it’s not entirely unexpected even though we are very disappointed at the number of unique problems that made it all the way to production.  Most embarrassing, we had problems in our new JavaScript agent that tracked client-side problems that prevented us from knowing about many issues customers were having which in turn gave us a false comfort in how well things were running.  We’ve substantially improved it (and those improvements are available to you as part of the Loupe Agent for JavaScript) and made significant progress in January and February identifying and resolving runtime problems folks were having.

What’s Next?

We’re currently reviewing the feature requests we’ve gotten from everyone and other input to set a product roadmap for the rest of 2016.  We’ll post all about it here in the blog when it’s ready (should be in March).

Rock solid centralized logging

Unlimited applications, unlimited errors, scalable from solo startup to enterprise.