All notable changes to this project will be documented in this file.
The format is based on Keep a Changelog and this project adheres to Semantic Versioning.
Unreleased section should follow Release Toolkit
- Add FIPS compliant packages
- Updated golang patch version to v1.23.5
- Updated golang patch version to v1.23.4
- Upgrade go to 1.23.2
- Upgrade integrations SDK so the interval is variable and allows intervals up to 5 minutes
- Updated golang version to v1.23.1
- Updated golang version to v1.22.5
- we were logging an error message even if there was no actual error causing a weird
\u003cnil\u003eto show up in the logs.
- Updated golang version to v1.22.3
- Updated golang version to v1.22.2
- Updated golang to version v1.21.7 to fix a vulnerability
- Updated github.com/newrelic/infra-integrations-sdk to v3.8.2+incompatible
- Updated golang version to 1.21
- Updated golang to v1.20.7
- bumped golang version pinning 1.20.6
- Update Go version to 1.20
- Exposed the 'QueuesMaxLimit' flag to configure the max amount of queues collected.
- Added a 'DisableEntities' flag to avoid generating inventory entries on large environments
- Fix log example file path in unix tarballs
- Fix: handle huge values in node's disk_free metric properly
- Disable CGO
- Upgrade Go to 1.19 and bump dependencies
- Add an arguments that allow to set the timeout to connecto to Rabbit's API
- Remove old Healthcheck and use the 'running' metric to calculate node status event.
- Logging configuration examples files.
- Removing the Cluster entity generation. This entity didn't contain any metrics related. And its deprecation was announced on this EOL
- Bump dependencies
Added support for more distributions:
- RHEL(EL) 9
- Ubuntu 22.04
- Move tool deps to go.mod in tools
- Update pipeline to Go 1.18
-
Replace the attribute
clusterNametorabbitmqClusterNameto avoid collisions with theclusterNameattribute reported when running in k8s. Naming was taken from HAproxy integration fix . User that have been useclusterNameattribute will need to replace it withrabbitmqClusterName. -
Adds the
host:portto all entity keys in order to use the entityRewrite when running in k8s. This fixes #73 . When the new version of the integration is deployed entities will be recreated with this new name. Old entities will live for an extra day. example:entityKey(before):ra-queue:/aliveness-test:clustername=rabbit@rabbitmq-0.rabbitmq-headless.rabbitmq.svc.cluster.localentityKey(fixed):ra-queue:k8s:k8s-cluster-name:rabbitmq:pod:rabbitmq-0:rabbitmq:15672:/aliveness-test:clustername=rabbit@rabbitmq-0.rabbitmq-headless.rabbitmq.svc.cluster.local
rabbitmq-log.yml.exampleis now in Linux packages to help setting up log parsing.
Added support for more distributions:
- Debian 11
- Ubuntu 20.10
- Ubuntu 21.04
- SUSE 12.15
- SUSE 15.1
- SUSE 15.2
- SUSE 15.3
- Oracle Linux 7
- Oracle Linux 8
Moved default config.sample to V4, added a dependency for infra-agent version 1.20.0
Please notice that old V3 configuration format is deprecated, but still supported.
- ARM support
- Update Go to v1.16.
- Migrate to Go Modules
- Update Infrastracture SDK to v3.6.7.
- Update other dependecies.
- ARM and ARM64 packages for Linux
- Exit rather than warn on API errors to avoid nil pointer error
- Moved release pipeline to Github Actions
- Increased queue limit to 2000
- Added argument
management_path_prefixto support custom prefix for all HTTP request to the rabbitmq management plugin as detailed here.
- Send an inventory value when it would otherwise be empty
- Renamed the integration executable from nr-rabbitmq to nri-rabbitmq in order to be consistent with the package naming. Important Note: if you have any security module rules (eg. SELinux), alerts or automation that depends on the name of this binary, these will have to be updated.
- Exclude windows definition from linux build
- Windows installer GUIDs
- Windows build files
- Re-added clusterName as a queryable value
- Exit code 69 error for rabbitmqctl
- Segfault regression
- Changed entity keys so they are more likely unique
- Updated to v3 SDK
- Added reportingEntity attribute
- Collect nodes when vhosts are filtered
- Fixed bug where Queue whitelist would not work
- Queue limiting was happening against full Queue list rather than filtered list
- Updated protocol version
- Bumped version to 1.0.0
- Increased Queue limit to 500
- Added limiting of Queue entities
- Changed sample file to be clearer for users to configure
- Correct misspellings
- Exchange Binding Metric Data would show up as Queue Metric Data. While the bug showed exchange bindings as queue bindings, it would always show a count of zero too
- Initial version: Includes Metrics, Inventory, and Events data