Docs

Outlyer Agent Changelog

v1.5.11

  • Add retries when check fails because of busy plugin

v1.5.10

  • Update check result to only return warning if check fails because the plugin is busy

v1.5.9

  • Capture log for exceptions in embedded check
  • Add remote instance hostname sanitization

v1.5.5

  • fix validation logic to prevent rpc trampling remote check validation

v1.5.4

  • update hostname lookup to use platform.node() before then socket.gethostname()

v1.5.3

  • pip outlyer_plugin extracted to allow plugin development outside the agent
  • Sanitize hostnames
  • Use runlevels in upstart config to ensure automatic starts

v1.5.2

  • Graceful shutdown flag

v1.5.1

  • Update platform detection in system_check
  • Add Sample method to Plugin API
    • The sample method is for recording a sample at a specific point in time without aggregation

v1.4.18

  • bugfix: ignore secondary IPs on NICs
  • enhancement: pykafka pip
  • Chocolately packaging for Windows installations

v1.4.15

  • ignore secondary IPs on NIC metrics

v1.4.14

  • collect metrics for all ‘up’ Linux NICs
  • add interface dimension to docker container NIC metrics

v1.4.13

  • Upgrade error from requests library

v1.4.12

  • Upgrade websockets client to 0.53
  • Update prometheus handler to support valid metric names
  • limit host labels to 256 chars

v1.4.6

  • Fix pass env variables to Upstart forked process
  • ntplib python pip for integrations

v1.4.5

  • Clearer proxy configuration
  • Fix for container PIDs
  • Internal python pip upgrades
  • Support for selectors on discovery providers
  • Upgrade cacerts
  • Upstart now runs as outlyer user

v1.3.5

  • Fix mixed-case metric labels for devices
  • Fix logging for embedded checks
  • Support IBM Cloud metadata
  • Publish metric TTL for step normalisation
  • Clear expired metrics from internal metric registry

v1.2.10

  • Fix loading python modules on Windows
  • Add embedded check log stream to check results
  • Support dots in command macros
  • Return critical status when an exception is generate in native plugins

v1.1.0

  • Publish check results

v1.0.19

  • embedded sqlite and python libs
  • restart of agent on upgrade

v1.0.18

  • pip dnspython 1.15.0
  • pip pymysql for Windows
  • Handle deregistration on service stop
  • Update host check ttl be interval + timeout
  • Remove requirement for ;warn;crit;min;max; in nagios perf data
  • Accept nagios perf data in host checks
  • example config updated with new options

v1.0.12

  • Added k8s.cluster label to all Kubernetes nodes so users can differentiate clusters in their account
  • Fixed devices being shown as containers when running RPC for plugin checks

v1.0.11

  • Add agent uptime metric agent.uptime to core metrics showing time in seconds since agent started
  • Added additional logging for failed checks in agent log
  • General performance improvments for agents running lots of checks under load

v1.0.10

  • additional debug logging for remote discovery and non-zero status
  • stderr and stdout in results page
  • ensure disabled checks are not scheduled

v1.0.8

  • fix for counter types in plugin api
  • fix for looking up instance parent

v1.0.6

  • sanitize metrics on instance creation
  • remove circus from the agent process
  • run the agent under systemd or upstart on Linux
  • enter mypy and typing for validations
  • apply host labels to containers

v0.1.11

  • for docker host use localhost ip if no httproxy
  • use container short-id for hostname
  • dont filter redis containers
  • add multiple stores for remote instance discovery
  • ensure system labels are lowercase

v0.1.8

  • pass instance labels as the instance environment
  • fix unnecessary check updating
  • ensure plugins latest plugins are loaded for active checks
  • sync plugins to plugin_cache
  • filter the kublet container
  • [pip ] upgrade boto3 1.7.26

v0.1.3

  • Start the changelog