DataHUB IQ 19.06.14 Release Notes

Executive Summary

We've launched the beginnings of several exciting new features this month:

1. Custom Metrics - create your own custom metrics using a set of simple operations on one or two other metrics.
2. Object Violation Dashboard - visualize network health at a glance on a per monitored object basis, sliced using whatever metadata makes sense to you (regions, network technology, etc)
3. Metadata in the TopN list - for when you have a hunch and want to validate it quickly from the dashboard without going into the compare workflow.
4. Metric Baselines and Dynamic Thresholds - giving you a new, more tailored way to determine when your network is misbehaving.

We're looking forward to getting your feedback and iteratively building them out a little more over the coming months.

Feature Highlights

Custom Metrics

Add, subtract, multiply or divide two metrics to create a custom metric. These custom metrics can then be visualized in object detail, object comparison, or dashboard views! Select the same metric twice with an add formula to generate a calculated round trip when dealing with two way measurement objects.

Object Violation Dashboard

We're listening to your feedback, and you've asked for more purpose built / pre-canned dashboard capabilities to augment the custom ones you know and love. The first of such purpose built dashboards is the Object Violations - visualize how much time your monitored objects spend violating their thresholds to get a true sense of network health.

Metadata in the TopN list

Sometimes you just want to know if there's a quick correlation between your TopN objects. Now that we've added the ability to add metadata columns, this correlation can be done a little faster than selecting all and issuing a comparison.

We've also added the ability to control how far the TopN list eats into the chart, and a way to temporarily hide the trend line.

Metric Baselines and Dynamic Thresholds

Ok so this one has already been launched for a little while now, but we didn't really announce it cause we wanted it to start working with your data in the background for a while. Here's the gist of the feature:

• Metric baselines - DataHUB is now maintaining a baseline for every metric of each monitored object. The baseline is a running average of what a particular metric looks like for a particular hour of the week (i.e. 8pm on Wednesdays). You can visualize this baseline in the object comparison view (look for the squiggly line icon beside the metric).
• Dynamic Thresholds - You can now generate threshold events when a monitored object deviates too far away from their baseline for a particular metric (eg. 25% or 50ms more than the baseline). This is another powerful tool that can help alert you of issues that go undetected with static, typically high water mark, thresholds.

Check here for more info on event thresholds

The detailed feature summary:

• Dashboards - titles can now be edited
• Dashboards - metadata can now be added to TopN lists
• Dashboards - object violation dashboard helps report on object health across metadata slices
• Custom metrics - add, subtract, multiply and divide two metrics to create a custom metric
• Object comparison - viewable metric baselines
• Interval comparison report - Option to see raw number or percentage deltas
• Threshold profile - dynamic thresholds based on metric baseline can be defined as an alternative to static levels
• Threshold profile - profiles can now be deleted

The detailed fix summary:

• Dashboards - Max value calculation was not consistent in chart when switching time period ranges
• Dashboards - zooming a time series on dashboard cards results in 0 value avg/min/max
• Dashboards - User role is not able to visualize any cards in dashboards
• Dashboards - inclusion vs exclusion metadata filters aren't easily differentiated
• Object detail - no direction (P2R/R2B/RT) for DMM metrics - show up as 0,1,2 in UI
• Object comparison - charts in metrics arrangement view (not clustered) can have their time series go out of sync
• Object comparison - charts fail to load when object list s is reduced
• Object comparison - normalization baseline causes Object comparison nothing showing in graphs
• Object comparison - don't force normalization when the chart arrangement is per-metric
• Object comparison - series aren't being removed when the associated objects are removed from comparison
• Object comparison - spark lines now showing in the comparison window
• Object comparison - metric selection becomes slow when selecting/deselecting a number of series in a short time span
• Object comparison - Comparison charts don't resolve metric graphing when toggling data cleaning off/on
• Interval comparison report - Line wrap of session names required
• Network - filter by object name is not working properly
• Network - filter by metadata is not working for network and monitor object view
• Network - support for "_ " in search terms
• Accounts - list of accounts not populating completely
• Data ingestion - add periods lost and lost burst to the packets section of the dictionary
• Data ingestion - object id definition for SOAM-DMM needs to guarantee uniqueness