Add help about how to get metrics

Added a flag --no-labels to skip printing labels in the output from l…

Add comment about external_labels

docs: link to grafana documentation for loki provisioning

add miss promtail client config option

Signed-off-by: Xiang Dai <764524258@qq.com>

Signed-off-by: Xiang Dai <764524258@qq.com>

Signed-off-by: Tom Wilkie <tom.wilkie@gmail.com>

Signed-off-by: Xiang Dai <764524258@qq.com>

Fix helm

Helm: Custom promtail configuration

Dropping liveness/readiness probe from promtail helm template as there is not health checks currently.. Can add back once something exist and we can do so without if statements..

Signed-off-by: Steven Sheehy <ssheehy@firescope.com>

Helm: Run with reduced privileges

Signed-off-by: Steven Sheehy <ssheehy@firescope.com>

Fix unused parameter warning for Fluentd

Add backoff_config in helm

Signed-off-by: Xiang Dai <764524258@qq.com>

Signed-off-by: Xiang Dai <764524258@qq.com>

Signed-off-by: Xiang Dai <764524258@qq.com>

Signed-off-by: Tom Wilkie <tom.wilkie@gmail.com>

Signed-off-by: Tom Wilkie <tom.wilkie@gmail.com>

Signed-off-by: Tom Wilkie <tom.wilkie@gmail.com>

Signed-off-by: Tom Wilkie <tom.wilkie@gmail.com>

* Adding the metric promtail_file_bytes_total which will report on the total size of each file with an active tailer.
Added an alert which will fire whenever the total bytes for a file is 100kb or more ahead the read bytes for more than 15mins.
Added an additional alert to track client bytes ready to send (encoded bytes) vs actually sent so we can add an alert on promtail send failures and still track stats on encoded bytes being sent.

* Moving where we update the total file bytes outside the tailer so it shouldn't get hung up by tailer issues.

* Recording file size based on Glob matches rather than what we are tailing to give even better insight

* Add a scrape config for pods without name or app labels.

Signed-off-by: Tom Wilkie <tom.wilkie@gmail.com>

* rebase with master and update to prefer name label but fallback to controller name

* Update scrape_config to use pod controller_name in the job

If a pod has a "name" label we use that in the job name. Otherwise,
- If the pod was created directly by a controller (statefulset,
  daemonset, job) then we use that controller_name.
- if the pod was created by controller created by a controller
  (deployment, cronJob) then we use the parent controller name.

- add a scrape_config.sh tool that generates the scrap_configs
  for promtail.sh and helm chart from the ksonnet lib.

* Reduce duplication in jsonnet.

Signed-off-by: Tom Wilkie <tom.wilkie@gmail.com>

* Fix rebase

Signed-off-by: Tom Wilkie <tom.wilkie@gmail.com>

* Tidy up promtail config imports.

Signed-off-by: Tom Wilkie <tom.wilkie@gmail.com>

Add readiness probe to Helm chart and cleanup

Signed-off-by: Steven Sheehy <ssheehy@firescope.com>

* Tag with latest


Signed-off-by: Xiang Dai <764524258@qq.com>