GitLab

Stop exporing per-backend nginx metrics

See merge request !132

nginx-related mtail metrics are quite high cardinality, the backend can
be discovered via logs when necessary.

bind-mount /etc/kibana directory within the container

See merge request !131

Remove all stretch-related configuration switches

Closes #78

See merge request !130

No need to support those anymore. Fixes issue #78.

Ship the seccomp.json file from containers-golang 0.3.2-87
and install it over the one provided in the 0.6.0 Debian package.

prometheus: expose 'job' in apache dashboard

See merge request !129

A service can have multiple apache containers, each in his own job.

This makes kernel messages appear again in the logs, the reason
though is still obscure.

Log failed jobs (non-zero exit status).

log-collector: disable kibana telemetry

See merge request !128

Let Grafana talk to all Thanos query

See merge request !127

Setting queue.minDequeueBatchSize as per
https://www.rsyslog.com/doc/v8-stable/rainerscript/queue_parameters.html
should reduce the number of queries we make to ES. Entries will be
as much as 3 seconds late.

Also sets max concurrency of the 'omstats' action to 1.

Alternative, more readable container run script

See merge request !126