v5.4.2

• filesystems monitor: Add mode dimension when new sendModeDimension option is set
• collectd/redis monitor: Add a few new metrics (thanks @xp-1000)
• Make bundled config files support APM input/output by default (will not affect existing deployments, only impacts APM PG deployments when using the bundled config with no modifications)

Docker Image: quay.io/signalfx/signalfx-agent:5.4.2 (digest: sha256:cc843c3140fd44720af661dd979f1dfe0c6813e7225f75846de04a1185e9c8d3)

v5.4.1

• jmx monitor: groovyScript option is now required (it always logically was before, just not enforced at startup)
• General improvements to trace correlation client error/retry handling
• diskio monitor: Added disk_ops.pending metric
• Added new montior telegraf/ntpq
• Added new option -service bypass to the Windows executable to facilitate running in a Windows container
• Fix bug where spans without any localEndpoint could cause a panic when writing

Docker Image: quay.io/signalfx/signalfx-agent:5.4.1 (digest: sha256:dff437d30fb5d075d74ac651ca42bfe93aa207bcf5374ed3f68c174902f7a1d4)

v5.4.0

• sql montior: Add ability to use expressions to generate datapoints from SQL queries
• Added new process monitor for montioring cpu and memory usage of individual processes
• kubelet-metrics montior: added container_image dimension when using /pods endpoint
• kubelet-metrics montior: Add timestamps to all metrics emitted
• Upgrade to newer version of collectd fork in order to get disk metrics on newer Linux kernel versions
• processlist montior: Fix potential token disclosure on Windows
• Fix nil panic that could happen when spans lack a localEndpoint object
• Add new metricNameTransformations config to monitor config that allows changing metric names as they come out of monitors

Docker Image: quay.io/signalfx/signalfx-agent:5.4.0 (digest: sha256:741beaa94376dbab683ea67267804b24921addd65bdf2e1a1dbdb8bd180d5c12)

v5.3.3

• collectd/mysql monitor: Add option to enable InnoDB stats (thanks @cvauvarin)
• statsd monitor: Fix race in metric handling that could cause a panic

Docker Image: quay.io/signalfx/signalfx-agent:5.3.3 (digest: sha256:f074329190a2fbe86a592b70dd62571e81d87a184c1bb3630337f84bc13670ca)

v5.3.2

• collectd/rabbitmq monitor: Add HTTPS support
• haproxy monitor: Add haproxy_session_limit and haproxy_session_max metrics (thanks @cvauvarin)
• processlist monitor: Handle double quotes in process command or username by changing them to single quotes
• kubernetes-volume monitor: Add more expressive dimensions for GCE disks and add volume type for standard volumes

Docker Image: quay.io/signalfx/signalfx-agent:5.3.2 (digest: sha256:e5b9fed04aec65a9d86f17a3da388225604259f0a7adf92b44ef9e3798e568c6)

v5.3.1

• Fix broken GCP host id dimension derivation (thanks @pdecat)
• Add span output to Splunk writer
• Automatically include kubernetes-cluster node condition metrics when configured so they are sent without having to add them to extraMetrics

Docker Image: quay.io/signalfx/signalfx-agent:5.3.1 (digest: sha256:2c7c173cd346b3f75c494e3c9d805aed619d4b0d17ebb50c27933bcd39d9ffc6)

v4.20.3

• Fix broken GCP host id dimension derivation (thanks @pdecat)

Docker Image: quay.io/signalfx/signalfx-agent:4.20.3 (digest: sha256:1b792d64d342fb86ae0cace5764074f9d7e9b5f65cc2b21cc9cbe31dd334d875)

v5.3.0

• Added new Splunk HEC output for datapoints and events (spans not yet supported). This can be enabled by setting the writer.splunk.enabled config to true and configuring the appropriate config options (see config docs)
• Add new kubelet-metrics monitor that is now favored over kubelet-stats due to the deprecation of the endpoints that kubelet-stats relied on. The new monitor does not get as many metrics, but will be available by default in newer K8s versions. The old monitor will still work the same if you have an existing configuration that relies on it.
• Add better endpoint configuration options to collectd/php-fpm and supervisor monitors (Thanks to @xp-1000)
• New mongodb-atlas monitor for monitoring remote Mongo Atlas clusters
• Made kubernetes.container_memory_limit and kubernetes.container_cpu_limit metrics default in the kubernetes-cluster monitor
• Added memory.total metric to memory monitor and made it default.
• Added cpu.num_processors metric to cpu monitor and made it default.
• Added discovered target dimensions to monitor status output

Docker Image: quay.io/signalfx/signalfx-agent:5.3.0 (digest: sha256:1ed6762162635e8ed72668030b84b32b036445fd7bb30dd968fc7e26ef995420)

v5.2.1

• Add telegraf/dns monitor to monitor DNS entries

Docker Image: quay.io/signalfx/signalfx-agent:5.2.1 (digest: sha256:fea0c76941d6a114edfe4f6bece340376d63a93a448827efaefcbb1b46c1a4d8)

v5.2.0

• Add additional inode-specific metrics to kubernetes-volumes monitor
• Add basic HTTP healthcheck monitor (http)
• Add cgroup monitor to directly monitor cgroup cpu and memory stats
• Updates to kubelet-stats monitor and k8s-api observer to better support EKS on Fargate
• Add first-class monitor for php-fpm based on collectd
• Add two more datapoints about the "updated" status of Kubernetes DaemonSets and Deployments
• Add a monitor for supervisor
• Updates to vsphere monitor for better efficiency in big clusters

Docker Image: quay.io/signalfx/signalfx-agent:5.2.0 (digest: sha256:328336de8454df582a3d41d955b156d36a628152deaf963e99785fc1fd2c10d3)