While this should always be the first service to be instantiated in a new environment, and for the most part it will work just fine since most of the resources are defined in such a way as to be independent of other services to avoid race-conditions. However, there are a few resources that reply on the outputs from other services, as such they cannot be created without causing a lovely chicken-and-egg scenario for the order in which to deploy said services.
To make life significantly easier any resources that rely on sub-services can be disabled by changing a single variable to true, while
this does force you to deploy this repo twice for a new cloud, that in significantly less effort than any alternatives. The variable can
even be set via the CI env so you don't have to commit any changes, just run the pipeline twice.
| Default value | TF Var | Env Var |
|---|---|---|
false |
first_deploy |
TF_VAR_first_deploy |
Ontop of the normal AWS credentials, this service requires some additional credentials to be provided.
| Name | Description |
|---|---|
| ELASTICSEARCH_URL | Elasticsearch URL |
| ELASTICSEARCH_USERNAME | Username for connecting to elasticsearch using basic auth |
| ELASTICSEARCH_PASSWORD | Password for connecting to elasticsearch using basic auth |