)]}'
{
  "log": [
    {
      "commit": "4fdb8e9d0ba398dad4e663482d1ce31e94b1f585",
      "tree": "56ab696cdbba2d2ec4a90f70e682284235747fb6",
      "parents": [
        "41a41d43c5cc5960ae0930c9332b62bc0446b77d"
      ],
      "author": {
        "name": "Aleš Komárek",
        "email": "github@newt.cz",
        "time": "Sat Nov 26 12:11:16 2016 +0100"
      },
      "committer": {
        "name": "GitHub",
        "email": "noreply@github.com",
        "time": "Sat Nov 26 12:11:16 2016 +0100"
      },
      "message": "Stacklight (#71)\n\n* Stacklight integration\r\n\r\n* Round 2\r\n\r\n* Variable service_name is missing for systemd file\r\n\r\n* preserve_data and ticker_interval are not strings\r\n\r\npreserve_data is a boolean, and ticker_interval is a number, so their values\r\nshouldn\u0027t have quotes.\r\n\r\n* Use \"ignore missing\" with the j2 include statement\r\n\r\n* Added cache dir\r\n\r\n* Use module_dir instead of module_directory\r\n\r\nThis fixes a bug where module_directory is used as the variable name instead of\r\nmodule_dir.\r\n\r\n* Use the proper module directory\r\n\r\nThe stacklight module dir is /usr/share/lma_collector/common, not\r\n/usr/share/lma_collector_modules. This fixes it.\r\n\r\n* Add the extra_fields.lua module\r\n\r\nThis commit adds the extra_fields Lua module. The extra fields table defined in\r\nthis module is empty right now. Eventually, this file will be a Jinja2 template\r\nand the content of the extra fields table will be generated based on the user\r\nconfiguration.\r\n\r\n* Regex encoder fix\r\n\r\n* Fix the decoder configuration\r\n\r\nThis commit uses proper decoder names in heka/meta/heka.yml. It also removes\r\nthe aggregator input for now, because it does not have an associated decoder.\r\n\r\n* Make Heka send metrics to InfluxDB\r\n\r\n* Add HTTP metrics filter to log_collector\r\n\r\n* Add logs counter filter to log_collector\r\n\r\n* Templatize extra_fields.lua file\r\n\r\n* Make InfluxDB time precision configurable\r\n\r\n* Configure Elasticsearch output through Pillar\r\n\r\n* Use influxdb_time_precision for InfluxDB output\r\n\r\nThis uses influxdb_time_precision set on metric_collector for configuring the\r\ntime precision in the InfluxDB output. This is to use just one parameter for\r\nboth the InfluxDB accumulator filter and InfluxDB output.\r\n\r\n* Increase maximum open files limit to 102400\r\n\r\n* Add alarming support\r\n\r\n* Revert \"[WIP] Add alarming support\"\r\n\r\n* Remove the aggregator output for now\r\n\r\nThis removes the aggregator output for now, as the aggregator doesn\u0027t work for\r\nnow. This is to avoid output errors in Heka.\r\n\r\n* Do not place Heka logs in /var/log/upstart\r\n\r\nWith this commit all the Heka logs are sent to /var/log/\u003cheka_service\u003e.log.\r\nPreviously, stdout was sent to /var/log/\u003cheka_service\u003e.log and stderr was sent\r\nto /var/log/upstart/\u003cheka_service\u003e.log, which was confusing to the operator.\r\n\r\n* Remove http check input plugin\r\n\r\nBecause it is not used anymore.\r\n\r\n* Add alarming support\r\n\r\n* Make the aggregator load heka/meta/heka.yml\r\n\r\nCurrently _service.sls does not load aggregator metadata from\r\nheka/meta/heka.yml. This commit fixes that.\r\n\r\n* Use filter_by to merge node grains data\r\n\r\n* Make the output/tcp.toml template extendable\r\n\r\n* Add an aggregator.toml output template\r\n\r\nThis template extends the tcp.toml output template.\r\n\r\n* Add generic timezone support to decoders\r\n\r\nThis change add a new parameter \u0027adjust_timezone\u0027 for the sandbox\r\ndecoder. This parameter should be set to true when the data to be\r\ndecoded doesn\u0027t contain the proper timezone information.\r\n\r\n* Add a run_lua_tests.sh script\r\n\r\nThis script will be used to run the Lua tests (yet to be added).\r\n\r\nTo run the script:\r\n\r\n    cd tests\r\n    ./run_lua_tests.sh\r\n\r\n* Copy Lua tests from fuel-plugin-lma-collector\r\n\r\n* Fix the afd tests\r\n\r\n* Fix the gse tests\r\n\r\n* Add aggregator config to support metadata\r\n\r\n* Fix the definition of the remote_collector service\r\n\r\nThis change removes unneeded plugins and adds the ones that are\r\notherwise required.\r\n\r\n* Fix state dependency\r\n\r\n* Add monitoring of the Heka processes\r\n\r\n* Set influxdb_time_precision in aggregator class\r\n\r\n* Disable the heka service completely\r\n\r\nWithout this patch `service heka status` reports that the heka service is\r\nrunning. For example:\r\n\r\nroot@ctl01:/etc/init.d# /etc/init.d/heka status\r\n * hekad is running\r\n\r\n* Define the highest_severity policy\r\n\r\n* Generate the gse_policies Lua module\r\n\r\n* Generate gse topology module for each alarm cluster\r\n\r\n* Generate gse filter toml for each cluster alarm\r\n\r\n* Adapt GSE Lua code\r\n\r\n* Remove gse cluster_field parameter\r\n\r\nThis parameter is not needed anymore. Heka\u0027s message_matchers are now used to\r\nmatch input messages.\r\n\r\n* Support dimensions in gse metrics\r\n\r\n* Do not rely on pacemaker_local_resource_active\r\n\r\n* Define the majority_of_members policy\r\n\r\n* Define the availability_of_members policy\r\n\r\n* Configure outputs in support metadata\r\n\r\n* Fix bug in map.jinja\r\n\r\nFix a bug in map.jinja where the filter_by for the metric_collector modified\r\nthe influxdb_defaults dict re-used for the remote_collector. The filter_by\r\nfunction does deep merges, so some caution is required.\r\n\r\n* Cleaning useless default map keys\r\n\r\n* Make remote collector send only afd metrics to influx\r\n\r\n* Add aggregator output to remote collector\r\n\r\n* Extend collectd decoder to support vrrp metrics\r\n\r\n* Update map.jinja\r\n\r\n* Update collectd decoder to parse ntpd metrics\r\n\r\n* Redefine alerting property\r\n\r\nThe alerting property can be one of \u0027disabled\u0027, \u0027enabled\u0027 or\r\n\u0027enabled_with_notification\u0027\r\n\r\n* Fix the gse_policies structure\r\n\r\nThe structure of the generated gse_policies.lua file is not correct. This\r\ncommit fixes that.\r\n\r\n* Add Nagios output for metric_collector\r\n\r\nThe patch embeds the Lua sandbox encoder for Nagios.\r\n\r\n* Add Nagios output for the aggregator\r\n\r\n* Send only alarm-related data to mine\r\n\r\n* Fix the grains_for_mine function\r\n\r\n* Fix flake8 in heka_alarming.py\r\n\r\n* Configure Hekad poolsize by pillar data\r\n\r\nThe poolsize must be increased depending on the number of filters.\r\nTypically, the metric_collector on controller nodes and the aggregator on\r\nmonitoring node(s) should probably use poolsize\u003d200.\r\n\r\n* Make Heka service watch Lua dir\r\n\r\nIn this way the service will restart when the content of\r\n/usr/share/lma_collector changes.\r\n\r\n* Enable collection of notifications\r\n\r\n* Add missing hostname variable in GSE code\r\n\r\n* Add a log decoder for Galera\r\n\r\n* Simplify message matchers\r\n\r\nThis removes the \"Field[aggregator] \u003d\u003d NIL\" part in the Heka message matchers.\r\n\r\nWe used to use a scribbler decoder to tag input messages coming in through the\r\naggregator input. We now have a dedicated Heka \"aggregator\" instance, so this\r\nmechanism is not necessary anymore.\r\n\r\n* Update collectd decoder for nginx metrics\r\n\r\n* Return an err message when set_member_status fails\r\n\r\nWith this commit an explicit error message is displayed in the Heka logs when\r\nset_member_status fails because the cluster has \"group_by\" set to \"hostname\"\r\nand an input message with no \"hostname\" field is received.\r\n\r\nThis addresses a comment from @SwannCroiset in #51.\r\n\r\n* Add contrail log parsers\r\n\r\n* Fix the heka grains for the aggregator/remote_collector\r\n\r\nPreviously, the heka salt grains of the node running aggregator/remote_collector\r\nget all the metric_collector alarms from all nodes (/etc/salt/grains.d/heka).\r\nThe resulting mines data is then wrong for the monitoring node, while that\r\nsitutation fortunately has no impact regarding metric_collector alarm\r\nconfigurations, the Nagios service leverging mine data get a wrong list of\r\nalarms for the monitoring node.\r\n\r\nThis patch fixes the issue with minimal changes but it appears that the logic\r\nbehind _service.sls state is not optimal and become hard to understand.\r\nThis state is executed several times with different contexts for every heka\r\n\u0027server\u0027 types and is not indempotent, indeed the /etc/salt/grains.d/heka file\r\ncontent is different between \u0027local\u0027 servers (metric|log)_collector and\r\n\u0027remote\u0027 servers remote_collector|aggregator.\r\n\r\n* Fix issue in lma_alarm.lua template\r\n\r\n* Add a log decoder for GlusterFS\r\n\r\n* Fix collectd Lua decoder for system metrics\r\n\r\nThe regression has been introduced by 74ad71d41.\r\n\r\n* Update collectd decoder for disk metrics\r\n\r\nThe disk plugin shipping with the 5.5. version of collectd (installed on\r\nXenial) provides new metrics: disk_io_time and disk_weighted_io_time.\r\n\r\n* Use a dimension key for the Nagios host displaying alarm clusters\r\n\r\n* Add redis log parser\r\n\r\n* Add zookeeper log parser\r\n\r\n* Add cassandra log parser\r\n\r\n* Set actual swap_size in collectd decoder\r\n\r\nSalt does not create Swap-related grains, but the \"ps\" module has\r\na \"swap_memory\" function that can be used to get Swap data. This commit\r\nuses that function to set swap_size in the collectd decoder.\r\n\r\n* Send annotations to InfluxDB\r\n\r\n* Add ifmap log parser\r\n\r\n* Support remote_collector and aggregator in cluster\r\n\r\nWhen deployed in a cluster, the remote_collector and aggregator\r\nservices are only started when the node holds the virtual IP address.\r\n\r\n* Add an os_telemetry_collector service\r\n\r\nos_telemetry_collector implements reading of Сeilometer samples\r\nfrom RabbitMQ and pulling them to InfluxDB (samples) and\r\nElasticSearch (resources)\r\n\r\n* heka server role, backward compat\r\n"
    }
  ]
}