A built-in health event was added to catch inconsistencies between the desired number of instances and those that are currently running. When the number of running instances is higher than 50%, but less than 100%, of the desired number of instances, the event is a warning. If it’s 50% or lower, the event is an error. For example, this event triggers when there are repeated crashes of Garden containers preventing the Cloud Foundry Application reaching a healthy state.
The Incidents view was rebuilt to reflect the visual changes and concepts that were applied in other product areas. There are also some small changes and improvements to help you understand and quickly find events.
A call has both source and destination applications; service and endpoint. Until now, tags like
service.name in Analyze were applied to the destination of calls. This feature also allows filtering and grouping on the source.
Infrastructure entity tags such as
host.name can also be applied to both the source and destination of a call. However, this option is not available for call tags such as
call.tag, which are properties on the call itself and are independent of the source or destination.
Sample queries with this feature:
- What are the endpoints called by shop service? Query: filter by
service.name=shopon the source and group by
endpoint.nameon the destination.
- What are the hosts that make database calls? Query: filter by
call.type=DATABASEand group by
host.nameon the source.
- Golang runtime: the search and analysis tags
go.versionare exposed for searching your infrastructure entities, traces, and calls.
- MongoDB: calls to a MongoDB cluster are mapped to a service named after the cluster replica set name.
- Nginx: calls to Nginx can be filtered by Nginx technology.
- AWS RDS: enhanced metrics gathering for the CPU credit usage and CPU credit balance.
- Kubernetes 1.15 support.
- Kubernetes namespace dashboard now displays used requests and limits even when no resource quota is set.
- AWS RDS: fixed false positives for the CPU credit balance health event.
- Couchbase: fixed service mapping using cluster name and fallback to host name. Added endpoint mapping rule using the bucket name.
- Docker: fixed false positives for the memory usage health event.
- Fixes issue that when the
canConfigureApplicationspermission was granted, an application could only be selected for the scope of a custom event, alert or maintenance window.
- Application Perspectives - Error Messages: fixed spike-chart to show correct numbers that are in line with the total count.