- [**CloudWatch:** Display partial result in graph when max DP/call limit is reached]({{< relref "#cloudwatch-display-partial-result-in-graph-when-max-data-points-per-call-limit-is-reached" >}})
This release adds a new panel named `Stat`. This panel is designed to replace the current `Singlestat` as the primary way to show big single number panels along with a sparkline. This panel is of course building on our new panel infrastructure and option design. So you can use the new threshold UI and data links. It also supports the same repeating feature as the Gauge and Bar Gauge panels, meaning it will repeat a separate visualization for every series or row
For the panels Gauge, Bar Gauge, and Stat, you can now leave the min and max settings empty. Grafana will, in that case, calculate the min and max based on all the data.
This panel shows RSS feeds as news items in the default home dashboard for v6.6. Add it to your custom home dashboards to keep up-to-date with Grafana news, or switch the default RSS feed to one of your choice.
A top feature request for years is now finally here. All panels now support custom units. Type any text in the unit picker and select the `Custom: <your unit>` option. By default, the text will be used as a suffix unit. If you want a custom prefix, then type `prefix: <your unit> ` to make the custom unit appear before the value. If you want a custom SI unit (with auto SI suffixes) specify `si:Ups`. A value like 1000 will be rendered as `1 kUps`.
The Bar Gauge visualization has a new display option: `Unfilled`. This new option is enabled by default, so it will change how this visualization is displayed on old dashboards. If you prefer the old default -- in which an unfilled area is not shown and the value follows directly after -- you have to update the visualization settings.
In order to align with a [change in Chrome 80](https://www.chromestatus.com/feature/5088147346030592), a breaking change has been introduced to Grafana's [`cookie_samesite` setting]({{< relref "../installation/configuration.md#cookie-samesite" >}}). Grafana now properly renders cookies with the `SameSite=None` attribute when this setting is `none`. The previous behavior of `none` was to omit the `SameSite` attribute from cookies. Grafana will use the previous behavior when `cookie_samesite` is set to `disabled`.
We introduced the wrap-lines option for logs because as for some of our users feel it's more efficient to see one line per log message. The wrapped-line option is set as a default; the unwrapped setting results in horizontal scrolling.
After feedback from our community, we have decided to reintroduce a labels column. However, for better readability and usefulness, we have transformed it into a Unique labels column which includes only non-common labels. All common labels are displayed above.
Isolating a series from a big set of lines in a graph is important for drill-downs. That's why we have implemented the context tooltip in Explore, which allows you to copy data and labels from it to further refine the query.
Now you can install the enterprise edition from the APT and YUM repository. The following table shows the APT repository for each Grafana version (for instructions read the [installation notes]({{< relref "../installation/debian/#install-from-apt-repository" >}})) :
The following table shows the YUM repositories for each Grafana version (for instructions read the [installation notes]({{< relref "../installation/rpm/#install-from-yum-repository" >}})) :
We recommend all users to install the Enterprise Edition of Grafana, which can be seamlessly upgraded with a Grafana Enterprise [subscription](https://grafana.com/products/enterprise/?utm_source=grafana-install-page).
From now on it will be possible to utilize meta data label in group bys, filters and in the alias field. Unfortunaltey there's no API to retrieve all the labels, but the group by field dropdown comes with a pre-defined list of common system labels. User labels cannot be pre-defined, but it's possible to enter them manually in the group by field. If a meta data label, user label or system label, is included in the group by segment, it will be possible to create filters based on it and to expand its value on the alias field.
{{<docs-imageboximg="/img/docs/v66/metadatalabels.gif"max-width="800px"caption="Stackdriver meta labels">}}
## CloudWatch: Calculate period based on time range
When the period field was left blank in Grafana 6.5, it would default to 60 seconds. In case users issued queries with a large time span, there was a high risk that they would reach the 100,800 data points per request limit in the Get Metric Data (GMD) api. When the period field is left blank in Grafana 6.6, the period will be calculated automatically based on the time range. The formula that is used is `time range in seconds / 2000`, and then we snap to next higher value in an array of pre-defined periods `[60, 300, 900, 3600, 21600, 86400]`. This will reduce the risk for receiving a `Too many datapoints requested` error in the panel.
## CloudWatch: Display partial result in graph when max data points per call limit is reached
In case all queries in a GMD call are metric stat (not using math expressions), Grafana will paginate the response until all data points are received. But pagination is not supported in case a math expression is being used, so in that case it's not possible to receive more than 100,800 data points. Previously when that limit was reached, we only displayed an error message. In Grafana 6.6, we also display the 100,800 data points that were received in the graph.
## Notice about upcoming changes in backendSrv for plugin authors
In our mission to migrate away from AngularJS to React we have removed all AngularJS dependencies in the core data retrieval service `backendSrv`. This change is already in master and will be introduced in the next `major` Grafana release.
Removing the AngularJS dependencies in `backendSrv` has the unfortunate side effect of AngularJS digest no longer being triggered for any request made with `backendSrv`. Because of this, external plugins using `backendSrv` directly may suffer from strange behaviour in the UI.
To remedy this issue as a plugin author you need to trigger the digest after a direct call to `backendSrv`.
