Deis Workflow is no longer maintained. Please read the announcement for more detail. |
|
---|---|
09/07/2017 | Deis Workflow v2.18 final release before entering maintenance mode |
03/01/2018 | End of Workflow maintenance: critical patches no longer merged |
Hephy is a fork of Workflow that is actively developed and accepts code contributions. |
Deis (pronounced DAY-iss) is an open source PaaS that makes it easy to deploy and manage applications on your own servers. Deis builds on Kubernetes to provide a lightweight, Heroku-inspired workflow.
This is an centos7 based image for running fluentd. It is built for the purpose of running on a kubernetes cluster.
This work is based on the docker-fluentd and docker-fluentd-kubernetes images by the fabric8 team. This image is in with deis v2 to send all log data to the logger component.
By default we do not capture kubernetes system logs. However, it is possible to tell fluentd to capture those logs just by specifying a few new environment variables.
- CAPTURE_START_SCRIPT
- CAPTURE_DOCKER_LOG
- CAPTURE_ETCD_LOG
- CAPTURE_KUBELET_LOG
- CAPTURE_KUBE_API_LOG
- CAPTURE_CONTROLLER_LOG
- CAPTURE_SCHEDULER_LOG
Set a variable's value to a non-empty string such as "true" to capture that log. Make these changes to the tpl/deis-logger-fluentd-daemon.yaml file in the Workflow chart directory.
To turn off log collection of fluentd's own logs to avoid infinite loops set the following environment variable to a non-empty string value
- DROP_FLUENTD_LOGS
To turn off the deis output plugin set the following environment variable to a non-empty string value
- DISABLE_DEIS_OUTPUT
To turn off sending log or metrics data to nsq set the following environment variable to "false"
- SEND_LOGS_TO_NSQ
- SEND_METRICS_TO_NSQ
This means we will not capture data from the log stream and send it to NSQ for processing. This means you will disable application logs (deis logs
) and metrics generated from deis router.
This plugin is used to decorate all log entries with kubernetes metadata.
Allows fluentd to send log data to an elastic search cluster. You must specify an ELASTICSEARCH_HOST
environment variable for this plugin to work.
ELASTICSEARCH_HOST="some.host"
ELASTICSEARCH_SCHEME="http/https"
ELASTICSEARCH_PORT="9200"
ELASTICSEARCH_USER="username"
ELASTICSEARCH_PASSWORD="password"
ELASTICSEARCH_LOGSTASH_FORMAT="true/false"
- Creates indexes in the formatindex_prefix-YYYY.MM.DD
ELASTICSEARCH_TARGET_INDEX_KEY="kubernetes.namespace_name"
- Allows the index name to come from within the log message map. See example message format below. This allows the user to have an index per namespace, container name, or other dynamic value.ELASTICSEARCH_TARGET_TYPE_KEY="some.key"
- Allows the user to set _type to a custom value found in the map.ELASTICSEARCH_INCLUDE_TAG_KEY="true/false"
- Merge the fluentd tag back into the log message map.ELASTICSEARCH_INDEX_NAME="fluentd"
- Set the index name where all events will be sent.ELASTICSEARCH_LOGSTASH_PREFIX="logstash"
- Set the logstash prefix variable which is used when you want to use logstash format without specifyingELASTICSEARCH_TARGET_INDEX_KEY
.ELASTICSEARCH_TIME_KEY=""
- specify where the plugin can find the timestamp used for the@timestamp
fieldELASTICSEARCH_TIME_KEY_FORMAT=""
- specify the format ofELASTICSEARCH_TIME_KEY
ELASTICSEARCH_TIME_KEY_EXCLUDE_TIMESTAMP=""
- IfELASTICSEARCH_TIME_KEY
specified dont set ``@timestamp
This plugin allows fluentd
to send data to a remote syslog endpoint like papertrail. You can configure fluentd
to talk to multiple remote syslog endpoints by using the following scheme:
SYSLOG_HOST_1=some.host
SYSLOG_PORT_1=514
SYSLOG_HOST_2=some.other.host
SYSLOG_PORT_2=52232
You can also set SYSLOG_HOST
and SYSLOG_PORT
.
This plugin allows for fluentd
to send all log data to a sumologic endpoint. You can configure it using the following environment variables:
SUMOLOGIC_COLLECTOR_URL
SUMOLOGIC_ENDPOINT
SUMOLOGIC_HOST
SUMOLOGIC_PORT
: defaults to 80 (unlessIS_HTTPS
is set and then its 443)IS_HTTPS
This plugin allows for fluentd
to send all log data to a remote graylog endpoint. You can configure it using the following environment variables:
GELF_HOST=some.host
GELF_PORT=12201
GELF_PROTOCOL="udp/tcp"
GELF_TLS="true/false"
GELF_TLS_OPTIONS_CERT="-----BEGIN CERTIFICATE-----\n[...]\n-----END CERTIFICATE-----"
GELF_TLS_OPTIONS_KEY="-----BEGIN PRIVATE KEY-----\n[...]\n-----END PRIVATE KEY-----"
GELF_TLS_OPTIONS_ALL_CIPHERS="true/false"
GELF_TLS_OPTIONS_TLS_VERSION=":TLSv1/:TLSv1_1/:TLSv1_2"
GELF_TLS_OPTIONS_NO_DEFAULT_CA="true/false"
Deis output is a custom fluentd plugin that was written to forward data directly to deis components while filtering out data that we did not care about. We have 2 pieces of information we care about currently.
-
Logs from applications that are written to stdout within the container and the controller logs that represent actions against those applications. These logs are sent to an internal messaging system (NSQ) on a configurable topic. The logger component then reads those messages and stores the data in an ring buffer.
-
Metric data from the nginx based router. We take the log and parse out
request_time
,response_time
, andbytes_sent
. Each one of these metrics makes up a series that we will ultimately send to our InfluxDB system. Attached to each series is the host the data came from (where router is running) and the status code for that request.
The topics these messages are put on are configurable via environment variables.
NSQ_LOG_TOPIC
NSQ_METRIC_TOPIC
If you need something beyond the plugins that come pre-installed in the image, it is possible to set some environment variables to install and configure custom plugins as well.
To install a custom plugin, simply set a FLUENTD_PLUGIN_# environment variable. For multiple plugins simply increment the trailing number.
FLUENTD_PLUGIN_1=some-fluentd-plugin
To configure your custom plugins, use either the CUSTOM_STORE_# or CUSTOM_FILTER_# environment variables
CUSTOM_STORE_1="configuration text"
CUSTOM_FILTER_1="configuration text"
If you need the build tools available for installing your plugin, this can be enabled with another environment variable
INSTALL_BUILD_TOOLS="true"