Skip to content

Release Notes



  • New Tagging API launched: Tagging API allows users to programmatically add, update, or delete user defined tags on hosts.



  • 6/20/2019 Windows Collector Release 1.12.20 : Support Upgrade, Auto-upgrade, Reconfiguration, Tags, Bug Fixing
  • 6/19/2019 Collector Release Version 1.12.21 : Patch Release
  • 6/17/2019 Collector Release Version 1.12.20
  • Performance improvement, AOC Connectivity improvements
  • Stats aggregation time changed from 1sec to 10 sec for L4-L7 analysis.
  • Logging optimization, Core-dump suppression, Debugging and diagnostics enhancements
  • General bug fixing



  • Collector Release Version 1.12.16. Performance, logging improvements and bug fixes.



  • Support for Monitoring Windows Servers
  • Windows Overview Dashboard. Maps, queries and alerting on Windows Metrics.


  • Support for Windows Server 2008, Windows Server 2012 and Windows Server 2016

1.12.17 - 1/17/2019


  • Design updates for maps
  • Various bug fixes


  • Windows support
  • Support for ICMP Protocol in IPFIX Records

1.12.8 - 11/26/2018


  • Changed default analysis depth to 'layer7' with 20% sampling rate
  • Disabled ssl-split by default
  • Collectors will listen on all interfaces by default

1.12.6 - 10/31/2018


  • Single table for dashboard/map lists
  • Design updates for integration list and API Settings
  • Increased chart resolution
  • Various bug fixes

1.12.1 - 09/27/2018


  • Lightweight layer4 analysis mode -- This mode analyzes network flows up till layer4 (IP and Transport) only. Analysis depth can be controlled via the parameter EPOCH_ANALYSIS_DEPTH which can be set as either layer4 (default) or layer7. Mapping works in both layer4 and layer7 modes. The layer7 analysis will show app level info whereas layer4 analysis will show only bytes, packets, and sessions but at a lower overhead. If you want layer7, we recommend setting the sampling rate to 10% in production.
  • Configuration variables have been renamed from Netsil to Epoch
  • Installation, configuration and logging directories moved from /opt/netsil/, /etc/netsil/, /var/log/netsil/ to /opt/nutanix/epoch/, /etc/nutanix/epoch/, /var/log/nutanix/epoch/ respectively.
  • Auto-upgrade to the current version of collectors is not supported. Older collectors are still compatible but are not recommended. Please re-install the latest collectors.

1.12.0 - 09/14/2018


  • 1-hop neighbors and group by
  • Auto-grouping of internal (collector installed) nodes -- "layer4_port" and "layer7_resource" in map group bys.
  • Port based auto-grouping, see "layer4_port".


  • New and consistent information architecture for buttons, forms and side-bar menu
  • Application Keys: Can be generated and revoked from the UI. Application Key is tied to a user account, it can be used to authenticate against the rest API with full user privileges.


  • New integration dashboards
  • AWS EC2
  • Consul
  • Tomcat
  • RabbitMQ
  • HAProxy
  • Elasticsearch
  • Zookeeper
  • Enhanced integration dashboards
  • System
  • Docker
  • Kafka


  • AWS integration: Agentless collection of infrastucture metrics from AWS CloudWatch

1.10.4 - 04/26/2018

  • New UI reskin
  • Faster loading for dashboard lists
  • Emails along with usernames can be used when logging into the AOC

1.10.2 - 03/16/2018

  • Replaced map library with cytoscape
  • Some fixes in migration

1.10.1 - 03/05/2018

  • Out of the box dashboards for Marathon and Mesos
  • Out of the box alert to detect when a host goes offline
  • Various bug fixes

1.10 - 02/15/2018

  • Overview Page: An out-of-the-box summary of key metrics for hosts, containers, kubernetes and service interactions.
  • "No Data" Alerts: Option to evaluate and trigger alerts when no data is received for an expected metric
  • Events: Ability to setup alerts on common events such as contatiners or pods restarting frequently. Ability to overlay docker and kubernetes events on any metrics in the analytics sandbox.
  • Simplified Analytics Sandbox: The UX controls for the analytics sandbox are less crowded and simplified.

1.9.3 to 1.9.10 - 12/25/2017

  • RPCAP TCP Channel - The new TCP based packet capture data channel 10-20% more reliable and performant.
  • Infrastructure tagging of custom metrics: Slice and dice custom metrics with Kubernetes or container tags.
  • Various bug fixes

1.9.2 - 12/05/2017

  • Visited the AOC UI and was greeted with a blank page with an error message? Now you will be greeted with the UI page, as per usual.
  • Improvements in topology loading experience.
  • The UI top bar now include links that hopefully answer questions such as "how do I install a collector?" or "are my collectors working?".
  • UI and collector bug fixes and improvements.

1.9.0 - 11/16/2017

  • HTTP/2 and GRPC protocol support -- Slice and dice by HTTP/2 and GRPC fields like GRPC status message, HTTP/2 path (grpc_service, grpc_method and grpc_type), HTTP/2 scheme etc.
  • Simpler Alert Workflow -- Single Alerts instead of Alert Templates, Instances and Rules
  • Out of the box alerts for Memory, Disk and HTTP Error Rates. The alerts would highlight health issues with nodes on the Host and Pod maps.
  • Quick alert shortcut on charts in dashboard.
  • Copy series button in drill-down menu -- a convenient way to copy series text.
  • Various bug fixes and backend improvements.

1.8.2 - 11/03/2017

  • Collector health page now shows output of collector info command.
  • Support for Marathon App and Mesos Task tags -- Can be used for grouping in Maps as well as a filter/group by in Analytics.
  • Collector install packages (debian and rhel) available for more OS versions.

1.8.0 - 10/17/2017

  • Collector Health Page
    • View health of collectors at a glance. Health is based on last check-in time to AOC.
    • View configuration parameters, version info. and tags at each collector.
  • Faster Maps Algorithm
    • Maps load 5-10x faster than the last version!
    • View Remote API's with a special group by option
  • Performance improvements for group by's on pod_name, container_name and host_name. Docker and Kubernetes dashboards load 2-3x faster now.
  • Improved Notification Emails
  • Table view widget -- Visualize multiple metrics in a tabular fashion

1.7.3 - 10/06/2017

  • Performance improvements and bug fixes

1.7.2 - 09/23/2017

  • Bugfixes

1.7.1 - 09/22/2017

  • Various bugfixes

1.7.0 - 09/19/2017

  • Multi-tenancy support
  • Create organizations to get complete isolation of metrics and assets for each organization (dashboards, maps, alerts etc.)
    • Cluster configuration moved to new super user app -- open http port 8443 on AOC
  • Imporved time-range controls
  • Enable or disable integrations to unlock dashboards, maps and alerts for an integration.
  • Various bug fixes

NOTE: Collectors must be auto-upgraded or re-installed in order to work with this release

1.6.2 - 08/22/2017

  • Memory optimization on collectors -- 2x improvement
  • Better content in default dashboards
  • Pods Map in default maps
  • Various bugfixes

1.6.1 - 08/17/2017

  • Various bugfixes

1.6.0 - 08/15/2017


  • Pin button on nodes to quickly add to map as pinned group
  • See containers and pods even if no traffic is running.

Dashboards and Visualizations

  • Bar charts re-design -- Easier toread series names
  • Iframe widget -- Include any web-page as an Iframe widget in your dashboards!
  • Line charts legend shows Avg/Min/Max values for each series in the visualization
  • Quick toggle to switch line charts to table view showing Avg/Min/Max of each series
  • Merged Summary and Trends dashboards


  • Compression on collector data channels -- Collectors now consume 4-10x lesser outgoing bandwidth
  • Stream processor on collector listens on 2005 instead of 2003 to prevent port clash with AOC

1.5.4 - 08/08/2017

  • Various bugfixes

1.5.3 - 08/03/2017

  • Various bugfixes

1.5.2 - 07/24/2017

  • Support for autodiscovery and autoconfiguration of service integrations

1.5.1 - 07/16/2017

  • Various bugfixes

1.5.0 - 07/12/2017

  • SAML based authentication: Configure single sign-on via SAML in AOC.
  • Pie chart visualization: Visualize groups in a pie chart as an alternative to bar chart.
  • Clone individual charts: Now you can clone individual charts to the same or a different dashboard.
  • Private Dashboards and Maps: Dashboards and Maps can be made visible to only the creator.
  • Read only permissions: Dashboards, Maps and Alerts can be marked "read only" such that only the users with admin permission may edit them.
  • Simplified collector routing: All the http/https communication from collector to AOC now takes place on a single port. Also, collectors communicate on https by default.
  • Hyperlink to external resources from chart drilldown. Hyperlinks are generated from url templates configured in settings and templated based on series clicked in a chart.
  • Collectors run in local stream processing mode by default.
  • Known issues:
    • If you are installing collectors as a Linux package on Ubuntu 15.10 or greater, please run it in collector mode, as there is a known issue with running it in the default local stream processing mode.

1.4.5 - 06/18/2017

  • Time rollup aggregate in Analytics. Can be used for creating capacity view queries and smoothening curves based on avg/max/min within each time bucket.
  • Now you can Star your favorite Dashboards and Maps.
  • Map updates
    • Link metrics (check gear icon)
    • Hide links (check gear icon)
    • Updated map icon designs
    • Updated Zoom-in/Zoom-out for icons
  • Various bugfixes

1.4.4 - 06/12/2017

  • Various bugfixes

1.4.3 - 06/09/2017

  • Fixed tag collection for Kubernetes
  • Fixed unit notation (MB, GB) for some of the metrics

1.4.2 - 06/08/2017

  • Various bugfixes

1.4.1 - 06/05/2017

  • Various bugfixes

1.4.0 - 05/31/2017

  • Maps updates
    • New map icons: The new icons highlight important information on the nodes more clearly.
    • Improved map drilldowns: Metrics, alerts and instances are much easier to navigate in the new maps drilldown view.
    • Groups page: Add or remove dashboards and alerts from the Groups page.
  • New Protocol metrics fields
    • DynamoDB keys: Now you can slice and dice the DynamoDB protocol metrics by DynamoDB keys
    • HTTP request parameters: Now you can slice and dice the HTTP protocol metrics by HTTP request parameters
  • Automatic tag collection on Google Cloud Platform
  • Standalone Stream Processor (beta): Run stream processor as a standalone component
  • Collection with local stream processing (beta): Stream process locally at the collector

1.3.1 - 05/08/2017

  • Various bugfixes

1.3.0 - 05/07/2017

  • Improved Alerts UX
    • New Triggered Alerts page that shows alerts in the time window.
    • Individual groups for alert rules that use GroupBy clause show up on the new Triggered Alerts page as separate entries.
    • Separated Alerts Settings experience from Alert Status experience to improve user flow, especially for alert rules that use GroupBy clause.
    • Cleaner Alert Templates rule builder page.
  • New Kubernetes Integration
  • Topology layout: Topology Maps now use cleaner force based layout instead of hierarchical layout
  • Auto-sampling of traffic: Bandwidth utilization based dynamic sampling algorithm at collectors (beta).
    • New algorithm uses TCP for data channel and closely monitors queue utilization and adjusts sampling rate accordingly.
    • Requires port 3003 (TCP) to be opened on AOC instance.
  • Collectors (public) IP override: Collectors can now be made to connect on (public) IP addresses of AOC worker instances, overriding the built-in load-balancing logic that typically distributes collector connections on private IP addresses of AOC workers.
  • Improved SMTP settings to provide control over SSL/TLS behavior.

1.2.4 - 04/25/2017

  • Fixes authentication issues with SMTP configuration

1.2.3 - 04/24/2017

  • Fixes rendering issues on Mozilla Firefox
  • Various bug fixes

1.2.2 - 04/23/2017

  • TTwitter transport support for Thrift (beta)
  • SMTP now supports non-TLS servers
  • Various bug fixes

1.2.1 - 04/14/2017

  • Various bug fixes & improvements

1.2.0 - 04/10/2017

  • Better UX
    • Improved topology map controls for dynamic and pinned groups. Use the collapsible right sidebar to see advanced controls.
    • Cleaner maps and dashboard pages.
    • Breadcrumbs based navigation.
    • Units for infrastructure metrics.
    • Faster UI load times.
  • Compression support on collectors -- Epoch AOC now supports compression to help reduce bandwidth overhead. Please see the instructions for more information.
  • DynamoDB analysis -- With Epoch AOC, you can now better understand DynamoDB performance (e.g. hot keys) and error rates (e.g. lookup failures) in real-time!
  • Bug fixes & improvements -- Docker metrics for Kubernetes, Stream processor performance improvements, Dynamic granularity on timeseries line charts.

1.1.3 - 03/27/2017

  • Various bug fixes

1.1.2 - 03/21/2017

  • Added support for offline license activation

1.1.1 - 03/20/2017

  • Topology Maps
    • Multiple topology maps -- Checkout the new topology maps page to see pre-canned maps and maps created by other users. You can also use the topology sandbox to explore and create different logical and physical topology maps.
    • With the new grouping, filtering and subgrouping options, you can now carve out precise & layered topology maps. For instance you can create maps that show interactions between services within a region, show interactions between regions, filter out production vs staging environments, show interactions between Kubernetes pods within a namespace and so on.
  • Improved filters -- Dashboard and topology filters more accessible and easier to use.
  • Cassandra analysis -- Now you can use Epoch AOC to gather deep performance and health insights into CQL interactions!

1.1.0 - 02/28/2017

  • Alerting improvements
    • Webhooks support - Now you can POST alert events to any other services with our template driven Webhooks support.
    • Alerts API : We have completely revamped our Alerting API, allowing you to programmatically create, update and delete alerts.
    • Alert emails - Now you can add a custom description (template based) and provide more context to alert emails.
  • Flexible service grouping
    • We have made the experience around service grouping more intuitive by showing live preview of compute instances that will match the grouping rule.
    • Infrastructure tags — Now you can use infrastructure tags, in addition to service resources (e.g. http.uri) in the grouping rules.
  • Infrastructure tags
    • Custom tags — Now you can provide custom tags (keys or key:value pairs) to the collectors. These tags are then assigned to each metric that is collected on the instance. Documentation
    • Kubernetes — We now support automatic tag/label collection (e.g. Kubernetes pods).
  • Improved Search Experience
    • Quickly search your services in the topology, dashboards and alerts.
    • Press backspace to escape the local page search to search all global resources (e.g. instances, http.uri etc.) in the AOC.
  • Backups/Restore — Now you can backup your entire AOC cluster (time series, events, user data) in an S3 compatible datastore.
  • Redis analysis — Now you can use Epoch AOC to analyze Redis interactions in real-time and set alerts on Redis SLA!
  • Cleaner Data model — Documentation
  • Various bug fixes and optimizations
  • Known Issues
    • When the hostname of the AOC node is too long, part of our internal DNS system will fail and break AOC startup. Thus, try to keep your hostname within 32 characters.

1.0.9 - 02/07/2017

  • Improved Alerts:
    • You can now set alerts on GroupBy queries and alerts will trigger for each group. For instance, you can now alert on any API endpoint that has < 99% availability.
    • PagerDuty integration: You can now send alert notifications to PagerDuty!
    • Alerts API: You can now set alerts via APIs. We are currently working on the documentation and in the meantime you can contact us on for assistance.
    • You can export the Alert instance JSON specification and use it elsewhere, such as an API call.
  • Analytics have improved speed and an improved query builder UI.
    • Now you can save the sandboxed analytics query as an Alert or a chart in a Dashboard.
    • You can now export the JSON specification of the analytics query and use it in other places such as an API call.
  • Charts now use a multi-color palette to help easily identify different timeseries.
  • Collectors are now VPC subnet aware and prefer connecting to active AOC workers within the local subnet.
  • Manual installation process for collectors has been improved.
  • Various bug fixes – Memcached protocol decoding performance, collectors package cleanup on purge/uninstall, collectors packages are now versioned properly and so on.

1.0.8 - 01/23/2017

  • Improved Alerting experience:
    • Rules driven Alert Templates provide a scalable and an organized way of defining setting alerts. Set alerts by applying filters and thresholds on Alert Templates.
    • Alerts support global scope and services scope.
    • Visual cues in topology and the new service health tab helps quickly identify performance and reliability issues.
    • Improved data source names.
    • Stay tuned for Pagerduty, Slack and HTTP Post support for receiving alert notifications in our next release!
  • Memcache analysis – Observe Memcached interactions down to the command level. Track and alert on latency of individual commands, errors and so on!
  • SSLSplit encryption keys – AOC Configuration now allows the user to configure keys for encrypting traffic mirrored by SSLSplit collector.
  • Epoch now uses Segment to report on usage metrics for better support and customer success. Automatic usage reporting is an opt-in feature.
  • Built-in deep storage facility for increased durability of timeseries data - Epoch now allows for storing and backing up timeseries data within the AOC cluster through Ceph - an S3 compatible object store.
  • Improved Settings page layout.
  • Memory optimizations and bug fixes.

1.0.7 - 12/14/2016

  • Support for defining dashboards and alerts using the new powerful analytics builder. Express analytics like never before with the new analytics system. It allows you to construct complex expressions and mathematical transforms on multiple data sources at once.
  • Timeshift operator -- Ability to timeshift any metric to an equivalent time window in the past. Can be used to express rate of change ("alert me if rps of a service changes by 3x") or to find periodicity in data (compare today vs yesterday).
  • New error rate widgets in default dashboards, e.g. ratio of 4xx or 5xx http response codes vs all response codes. High error rates are a strong indicator for service health in production and is one of the four golden signals of API monitoring.
  • SSLsplit collector is now available as a stand-alone package, provisioned as a long running transparent proxy.
  • Various bug fixes - percentile metrics correctness, all operational assets such as JS and CSS files are bundled in the AOC installation etc.

1.0.6 - 11/28/2016

  • Service health tab -- The health tab provides a holistic view of your service health at a glance. The health is computed based on the imported alert bundles. The health compuation is compatible with time travel so you can view service health in the past.
  • Alert Bundles -- An alert bundle is a collection of alert templates which can be imported in a service just like a dashboard. The same alert bundle(s) can be imported across different services thus reducing the overhead of maintaining a unique alert policy for each service. The filters and thresholds for an alert bundle can be configured while importing.
  • Analytics Sandbox -- Express analytics like never before with the new analytics system. It allows you to construct complex expressions and mathematical transforms on multiple data sources at once. Note: Full dashboard and alert support for new analytics will be available in the next release.
  • PostgreSQL analysis -- Observe PostgreSQL interactions down to the query level. Track and alert on latency for individual queries, server errors and so on! Also includes PostgreSQL metrics integration (reported by PostgreSQL) in addition to interaction analysis.
  • Enhanced stream processor performance -- AOC can now analyze more interactions per second than ever before!

1.0.5 - 11/14/2016

Note: If you had containerized collectors or collectors installed on RHEL-based systems, please remove those collectors before upgrading Epoch AOC to 1.0.5. Collector auto-upgrades only worked for Debian-based systems in releases prior to 1.0.5.

  • Thrift analysis -- Observe thrift interactions down to the method names. Track and alert on latency for individual methods, response types and so on!
  • Verified services view -- All the verified services can now be viewed and edited at a single location in the configuration section.
  • Better cluster management -- Exposed DC/OS, Marathon and Zookeeper Exhibitor User Interface in the configuration page. You may have to authenticate at the endpoint https://your.epoch.url:4443 before accessing the cluster management page.
  • Auto-upgrades for RHEL-based and containerized collectors -- The container collectors can now auto-upgrade when Epoch AOC is upgraded.
  • Various bugfixes in the AOC upgrade system.

1.0.4 - 10/21/2016

Note: This release has multiple changes to the underlying platform and it is highly recommended that your existing installation is replaced with a fresh instance of Epoch AOC and its latest collectors.

  • Added ability to filter interaction metrics by server ports.
  • AWS EC2 instance tags: Now you can configure automatic AWS EC2 instance tags collection and segment your interaction metrics and infrastructure metrics based on them.
  • Collector auto-upgrades (for Linux collectors): Your existing collectors now auto-upgrade when Epoch AOC is upgraded. Stay tuned for auto-upgrades of Docker Container collectors in our next release.
  • Services Table View: In addition to services topology, now you can view your services in a tabular form as well for a quick overview of their health and metrics.
  • Various bug and stability fixes:
    • Downgraded Mesosphere DC/OS from version 1.8 to 1.7 to prevent volume removal on Epoch AOC instance reboot.
    • Exposed port 80 for HTTP access to Web UI (optional).
    • Metrics data auto-evicts when the disk is getting full.
    • No more hanging processes when collectors are re-downloaded and re-installed on bare-metal hosts. To fix the collectors on your current hosts:
      1. Stop the current collectors with /etc/init.d/epoch-collectors stop.
      2. Completely remove the epoch-collectors package from your system with sudo apt-get purge epoch-collectors or sudo yum remove epoch-collectors, depending on your operating system
      3. Run ps -eaf | grep collectors to identify any hanging processes. Note their PIDs and kill them with kill -9 <pid>
      4. Re-install the collectors from the Help page.
  • Single-container packages of Epoch AOC now come embedded with Druid time-series database for efficient long term data retention.

1.0.2 - 10/10/2016

  • DNS Analysis: Observe DNS interactions down to query level. Track DNS errors such as "Non-Existent Domain" and "Lookup Timeouts" and measure response latencies. Don't let DNS issues impact critical applications!
  • Service Link Metrics: In addition to tracking service-level metrics, AOC now lets you pin-point potential issues and hotspots between any pair of service. A great way to detect client errors (4xx), incorrect API invocations etc.

1.0.1 - 03/10/2016

  • Various stability fixes
  • Scalable topology clustering algorithm — Scales to large number of unique resources in a large and diverse application

1.0.0 - 09/25/2016

  • Production grade features:

    • Scalable Ingestion — Horizontally scale stream-processors to analyze web-scale apps.
    • Fast analytics — Powered by Druid, a timeseries optimized database (
    • Scalable Storage — Local mount and S3 are available as deep-storage options.
    • 1-click updates for the cluster.
    • Easy administration — Powered by DC/OS.
  • Attach dashboards to a service profile — allows user to build custom dashboard templates and load them with the context of a service. This alleviates the need to clone dashboards each time a new service is introduced.

  • MySQL metrics integration, in addition to MySQL interaction analytics.

  • Various bug fixes

0.2.36 - 09/12/2016

  • New charting library based on HighCharts.
  • Global search capabilities to help quickly locate services, instances and resources such as REST API endpoints.
  • Improved default dashboards — REST APIs, MySQL, NetworkFlows, DNS, System & Docker stats.
  • Filter templates on dashboards for faster orientation during incidents and planning. Quickly load dashboards with right data with a few clicks!
  • MySQL analysis — Observe MySQL interactions down to the query level. Track and alert on latency for individual. queries, server errors and so on!
  • NetworkFlow analysis — Track and alert on TCP connection performance metrics such as retransmissions, out of order counts and so on!
  • Various bug fixes and improvements.
  • Stay tuned for our next release (3rd week of Sept’16) which will have production grade features such as clustering capability to scale traffic processing, scalable time series database and deep storage options for S3 and HDFS.