Configuration to allow users to retry loading dashboards #6560
Conversation
| @@ -21,13 +23,13 @@ type KibanaLoader struct { | |||
| msgOutputter MessageOutputter | |||
| } | |||
|
|
|||
| func NewKibanaLoader(cfg *common.Config, dashboardsConfig *Config, hostname string, msgOutputter MessageOutputter) (*KibanaLoader, error) { | |||
| func NewKibanaLoader(ctx context.Context, cfg *common.Config, dashboardsConfig *Config, hostname string, msgOutputter MessageOutputter) (*KibanaLoader, error) { | |||
There was a problem hiding this comment.
exported function NewKibanaLoader should have comment or be unexported
libbeat/dashboards/config.go
Outdated
| Retry *Retry `config:retry` | ||
| } | ||
|
|
||
| type Retry struct { |
There was a problem hiding this comment.
exported type Retry should have comment or be unexported
jalvz
force-pushed
the
retry-load-dashboards
branch
2 times, most recently
from
2371159 to
4ac7c1a
Compare
libbeat/dashboards/config.go
Outdated
| } | ||
|
|
||
| var defaultConfig = Config{ | ||
| KibanaIndex: ".kibana", | ||
| AlwaysKibana: false, | ||
| Retry: &Retry{ | ||
| Enabled: false, |
There was a problem hiding this comment.
Can you add the defaults for interval and maximum here too?
There was a problem hiding this comment.
sure thing, what do you think is reasonable?
1 sec/ unlimited?
There was a problem hiding this comment.
SGTM, especially as the feature is off by default.
There was a problem hiding this comment.
done 3bd3affaa9242bd7072207a911f1313ead19c30e
jalvz
force-pushed
the
retry-load-dashboards
branch
2 times, most recently
from
4ae6de7 to
fdc0308
Compare
|
added changelog fdc0308f204f20d088b571460a509bae5b2427a5 |
There was a problem hiding this comment.
Could you add the config options to the config reference file here (https://github.com/elastic/beats/blob/master/libbeat/_meta/config.reference.yml#L639) and run make update in the beats directory?
After that we should be good to go.
| @@ -99,3 +99,19 @@ NOTE: This setting only works for Kibana 6.0 and newer. | |||
|
|
|||
| Force loading of dashboards using the Kibana API without querying Elasticsearch for the version | |||
| The default is `false`. | |||
|
|
|||
| [float] | |||
| ==== `setup.dashboards.retry.enabled` | |||
… not reachable, instead of exiting right away.
jalvz
force-pushed
the
retry-load-dashboards
branch
from
fdc0308 to
68d3c51
Compare
filebeat/filebeat.reference.yml
Outdated
| @@ -1223,6 +1223,17 @@ output.elasticsearch: | |||
| # how to install the dashboards by first querying Elasticsearch. | |||
| #setup.dashboards.always_kibana: false | |||
|
|
|||
| # If true and Kibana is not reachable at the time when dashboards are loaded, | |||
| # it will retry to reconnect to Kibana instead of exiting with an error. | |||
| #setup.dashboards.retry.enabled | |||
There was a problem hiding this comment.
If there are default values we normally put them into the reference doc, so here it would be #setup.dashboards.retry.enabled: false. Similar for the values below.
| [float] | ||
| ==== `setup.dashboards.retry.enabled` | ||
|
|
||
| If this option is set to true, and Kibana is not reachable at the time when dashboards are loaded, |
There was a problem hiding this comment.
My comment above made me realise that having a comment about the defaults here would also be great. Some at then of the sentence like. Default is false.
|
Thanks for the addition. |
Configuration to allow users to retry loading dashboards if Kibana is not reachable, instead of exiting right away.
... if Kibana not reachable, instead of exiting right away.