DNSMON Testing page

Last update: 2014-03-05

Introduction

This page about DNSMON was created specially for the group of testers during the introductory period and is therefore not meant to be shared with the general public, although some sections will probably be included in the documentation and general description of the service.

We encourage all testers to read through this page first. After this, if you feel brave enough, please go ahead and explore the new interface for yourself; it should be intuitive enough, especially for someone who is aware of the DNSMON concepts. However, we recommend reading a bit of documentation first, especially about the features of the visualiser. See below for more details.

Current state of the new DNSMON (and differences from the original)

Even though the basic idea for the DNSMON service remains the same, in the past months, we rebuilt the machinery on a new foundation. Instead of TTM, it now uses RIPE Atlas as the data collection mechanism. Specifically, the RIPE Atlas anchors are involved in measuring the reachability and responses of the name servers of the monitored zones. We've also completely rewritten the visualisation side, to modernise it and make it more interactive. These changes result in a couple of consequences:

• The set of vantage points is now different (RIPE Atlas anchors vs. TTM boxes).
• RIPE Atlas is in the process of installing more anchors every week. At the time of writing this, there are 37 anchors up and running, contributing to these measurements. We plan to have 50 by the end of Q1 2014. Although we also involved regular RIPE Atlas probes when we started these measurements, we've decided to stick to the more reliable anchors and will add them to the existing measurements as they come alive.
• The raw data format is different than the one used by RIPE Atlas. Also, this data is now accessed via the RIPE Atlas website and APIs.
• The new system does a lot more than the previous one did, for example:
  • Support for TCP queries in addition to UDP
  • Support for traceroutes towards the servers, and access to these via the interface
  • (Potential) support for visualising server instances of an anycasted service
  • The client side interactivity allows zooming in on interesting details (e.g. servers, vantage points, time intervals)
  • User-defined thresholds for visualising the observed error rates (i.e. what's red, what's green and how wide the interval is between these)

There are some features of DNSMON that we are not planning to migrate to the new implementation:

• Data delay for non-DNSMON users; the measurements done by RIPE Atlas for the new DNSMON are public measurements; therefore, the results are publicly available.
• Server-side generated RRD graphs; the client side visualisation is meant to provide a comparable replacement.
• The RIPE Atlas measurements are not retried on failure. The RIPE Atlas measurements are retried 10 times on failure. We will consider changing this, or changing how we visualize retries, based on discussions with the community.

When adding a new zone to be monitored in the new DNSMON, we look up all the name servers defined for that zone, and schedule RIPE Atlas DNS measurements against these. Specifically, the following measurements are defined (both IPv4 and IPv6, if possible):

¹ HOSTNAME.BIND and VERSION.BIND queries are used by default because of their widespread adoption. For servers that do not support these forms, ID.SERVER and VERSION.SERVER will be used instead.

² This frequency is less than the original DNSMON. We will consider making HOSTNAME.BIND queries more frequent based on the feedback we receive from DNSMON users and other interested parties.

³ This frequency is greater than the SOA measurements in the original DNSMON, although less than the frequency of HOSTNAME.BIND queries. This is because the visualisations in the new DNSMON are driven by the SOA data in order to capture interesting RCODEs (e.g. NXDOMAIN). We will consider making SOA measurements more frequent based on feedback.

Works in progress

We're still working on the following features:

• Bug fixes – see below for details on known and fixed bugs
• More/better documentation about the visualiser features and site navigation
• Change URL after user click to support back/forward in the browser
• Nicer URLs (also permalinks)
• Enhance (textual) content on various pages
• Easier/alternative navigation for zone (servers) -> server (probes) transitions NOFIX, the current navigation seems to be enough
• Migrate zones currently monitored by DNSMON
• (2014-02-12) Add a view for visualising relative RTT changes, in addition to the absolute values

Known bugs, issues and possible enhancements

The following bugs are known / have been reported / have been fixed already: 

• The "monitor" (auto-refresh) button is not functioning yet;
• (2014-02-11) The datepicker can not be opened twice.
• (2014-02-12) The threshold selector resets the values on invocation
• (2014-02-12) Users should be able to enter thresholds as values, not only by shifting sliders
• (2014-02-13) Make the visualised time interval limits appear as text
• (2014-02-13) Date picker should allow time selection in addition to date
• (2014-02-13) Add a button to "jump to latest data"
• (2014-02-13) Some links requiring Atlas login don't take you to the intended destination
• (2014-02-13) Add a button to "Download raw data for what I see now" NOFIX until there's enough user demand
• (2014-02-18) Add the server/probe and the timestamp to the cell popup
• (2014-02-20) Sometimes sharing a view with a permalink changes the data resolution
• (2014-02-21) Add "closest traceroute before/after this result" feature to the native resolution cell details + show parsed version of a result

Under consideration

There are some features that we're considering, or already plan to add at a later stage:

• Provide more direct links to the raw data in RIPE Atlas
• Clicking on the probe in a single probe view should take the user "somewhere useful", such as to the RIPE Atlas anchor page (at the moment it leads back to the same page)
• Probe grouping and smart ordering: the y-axis (in probe view) can use different orderings, such as country order, AS order...
• Allow full-screen operations
• Implement instance view visualisations
• Make an SVG export of the current view
• Move threshold changer next to the values on the left-hand side, and make all of this clickable in order to bring up the threshold selector
• Investigate visualising all servers that are seen from a particular vantage point

Documentation

Please see the the documentation page.

Contact us

If you are testing the new DNSMON and have feedback, please send an email to the DNSMON mailing list at dnsmon-test@ripe.net.